/*
* qemu_migration.c: QEMU migration handling
*
* Copyright (C) 2006-2015 Red Hat, Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see
* .
*
*/
#include
#include
#include
#include
#ifdef WITH_GNUTLS
# include
# include
#endif
#include
#include
#include "qemu_migration.h"
#include "qemu_monitor.h"
#include "qemu_domain.h"
#include "qemu_process.h"
#include "qemu_capabilities.h"
#include "qemu_command.h"
#include "qemu_cgroup.h"
#include "qemu_hotplug.h"
#include "qemu_blockjob.h"
#include "domain_audit.h"
#include "virlog.h"
#include "virerror.h"
#include "viralloc.h"
#include "virfile.h"
#include "virnetdevopenvswitch.h"
#include "datatypes.h"
#include "fdstream.h"
#include "viruuid.h"
#include "virtime.h"
#include "locking/domain_lock.h"
#include "rpc/virnetsocket.h"
#include "virstoragefile.h"
#include "viruri.h"
#include "virhook.h"
#include "virstring.h"
#include "virtypedparam.h"
#include "virprocess.h"
#include "nwfilter_conf.h"
#include "storage/storage_driver.h"
#define VIR_FROM_THIS VIR_FROM_QEMU
VIR_LOG_INIT("qemu.qemu_migration");
VIR_ENUM_IMPL(qemuMigrationJobPhase, QEMU_MIGRATION_PHASE_LAST,
"none",
"perform2",
"begin3",
"perform3",
"perform3_done",
"confirm3_cancelled",
"confirm3",
"prepare",
"finish2",
"finish3",
);
enum qemuMigrationCookieFlags {
QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS,
QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE,
QEMU_MIGRATION_COOKIE_FLAG_PERSISTENT,
QEMU_MIGRATION_COOKIE_FLAG_NETWORK,
QEMU_MIGRATION_COOKIE_FLAG_NBD,
QEMU_MIGRATION_COOKIE_FLAG_STATS,
QEMU_MIGRATION_COOKIE_FLAG_MEMORY_HOTPLUG,
QEMU_MIGRATION_COOKIE_FLAG_LAST
};
VIR_ENUM_DECL(qemuMigrationCookieFlag);
VIR_ENUM_IMPL(qemuMigrationCookieFlag,
QEMU_MIGRATION_COOKIE_FLAG_LAST,
"graphics",
"lockstate",
"persistent",
"network",
"nbd",
"statistics",
"memory-hotplug");
enum qemuMigrationCookieFeatures {
QEMU_MIGRATION_COOKIE_GRAPHICS = (1 << QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS),
QEMU_MIGRATION_COOKIE_LOCKSTATE = (1 << QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE),
QEMU_MIGRATION_COOKIE_PERSISTENT = (1 << QEMU_MIGRATION_COOKIE_FLAG_PERSISTENT),
QEMU_MIGRATION_COOKIE_NETWORK = (1 << QEMU_MIGRATION_COOKIE_FLAG_NETWORK),
QEMU_MIGRATION_COOKIE_NBD = (1 << QEMU_MIGRATION_COOKIE_FLAG_NBD),
QEMU_MIGRATION_COOKIE_STATS = (1 << QEMU_MIGRATION_COOKIE_FLAG_STATS),
QEMU_MIGRATION_COOKIE_MEMORY_HOTPLUG = (1 << QEMU_MIGRATION_COOKIE_FLAG_MEMORY_HOTPLUG),
};
typedef struct _qemuMigrationCookieGraphics qemuMigrationCookieGraphics;
typedef qemuMigrationCookieGraphics *qemuMigrationCookieGraphicsPtr;
struct _qemuMigrationCookieGraphics {
int type;
int port;
int tlsPort;
char *listen;
char *tlsSubject;
};
typedef struct _qemuMigrationCookieNetData qemuMigrationCookieNetData;
typedef qemuMigrationCookieNetData *qemuMigrationCookieNetDataPtr;
struct _qemuMigrationCookieNetData {
int vporttype; /* enum virNetDevVPortProfile */
/*
* Array of pointers to saved data. Each VIF will have its own
* data to transfer.
*/
char *portdata;
};
typedef struct _qemuMigrationCookieNetwork qemuMigrationCookieNetwork;
typedef qemuMigrationCookieNetwork *qemuMigrationCookieNetworkPtr;
struct _qemuMigrationCookieNetwork {
/* How many virtual NICs are we saving data for? */
int nnets;
qemuMigrationCookieNetDataPtr net;
};
typedef struct _qemuMigrationCookieNBD qemuMigrationCookieNBD;
typedef qemuMigrationCookieNBD *qemuMigrationCookieNBDPtr;
struct _qemuMigrationCookieNBD {
int port; /* on which port does NBD server listen for incoming data */
size_t ndisks; /* Number of items in @disk array */
struct {
char *target; /* Disk target */
unsigned long long capacity; /* And its capacity */
} *disks;
};
typedef struct _qemuMigrationCookie qemuMigrationCookie;
typedef qemuMigrationCookie *qemuMigrationCookiePtr;
struct _qemuMigrationCookie {
unsigned int flags;
unsigned int flagsMandatory;
/* Host properties */
unsigned char localHostuuid[VIR_UUID_BUFLEN];
unsigned char remoteHostuuid[VIR_UUID_BUFLEN];
char *localHostname;
char *remoteHostname;
/* Guest properties */
unsigned char uuid[VIR_UUID_BUFLEN];
char *name;
/* If (flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) */
char *lockState;
char *lockDriver;
/* If (flags & QEMU_MIGRATION_COOKIE_GRAPHICS) */
qemuMigrationCookieGraphicsPtr graphics;
/* If (flags & QEMU_MIGRATION_COOKIE_PERSISTENT) */
virDomainDefPtr persistent;
/* If (flags & QEMU_MIGRATION_COOKIE_NETWORK) */
qemuMigrationCookieNetworkPtr network;
/* If (flags & QEMU_MIGRATION_COOKIE_NBD) */
qemuMigrationCookieNBDPtr nbd;
/* If (flags & QEMU_MIGRATION_COOKIE_STATS) */
qemuDomainJobInfoPtr jobInfo;
};
static void qemuMigrationCookieGraphicsFree(qemuMigrationCookieGraphicsPtr grap)
{
if (!grap)
return;
VIR_FREE(grap->listen);
VIR_FREE(grap->tlsSubject);
VIR_FREE(grap);
}
static void
qemuMigrationCookieNetworkFree(qemuMigrationCookieNetworkPtr network)
{
size_t i;
if (!network)
return;
if (network->net) {
for (i = 0; i < network->nnets; i++)
VIR_FREE(network->net[i].portdata);
}
VIR_FREE(network->net);
VIR_FREE(network);
}
static void qemuMigrationCookieNBDFree(qemuMigrationCookieNBDPtr nbd)
{
if (!nbd)
return;
while (nbd->ndisks)
VIR_FREE(nbd->disks[--nbd->ndisks].target);
VIR_FREE(nbd->disks);
VIR_FREE(nbd);
}
static void qemuMigrationCookieFree(qemuMigrationCookiePtr mig)
{
if (!mig)
return;
qemuMigrationCookieGraphicsFree(mig->graphics);
qemuMigrationCookieNetworkFree(mig->network);
qemuMigrationCookieNBDFree(mig->nbd);
VIR_FREE(mig->localHostname);
VIR_FREE(mig->remoteHostname);
VIR_FREE(mig->name);
VIR_FREE(mig->lockState);
VIR_FREE(mig->lockDriver);
VIR_FREE(mig->jobInfo);
VIR_FREE(mig);
}
#ifdef WITH_GNUTLS
static char *
qemuDomainExtractTLSSubject(const char *certdir)
{
char *certfile = NULL;
char *subject = NULL;
char *pemdata = NULL;
gnutls_datum_t pemdatum;
gnutls_x509_crt_t cert;
int ret;
size_t subjectlen;
if (virAsprintf(&certfile, "%s/server-cert.pem", certdir) < 0)
goto error;
if (virFileReadAll(certfile, 8192, &pemdata) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unable to read server cert %s"), certfile);
goto error;
}
ret = gnutls_x509_crt_init(&cert);
if (ret < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot initialize cert object: %s"),
gnutls_strerror(ret));
goto error;
}
pemdatum.data = (unsigned char *)pemdata;
pemdatum.size = strlen(pemdata);
ret = gnutls_x509_crt_import(cert, &pemdatum, GNUTLS_X509_FMT_PEM);
if (ret < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot load cert data from %s: %s"),
certfile, gnutls_strerror(ret));
goto error;
}
subjectlen = 1024;
if (VIR_ALLOC_N(subject, subjectlen+1) < 0)
goto error;
gnutls_x509_crt_get_dn(cert, subject, &subjectlen);
subject[subjectlen] = '\0';
VIR_FREE(certfile);
VIR_FREE(pemdata);
return subject;
error:
VIR_FREE(certfile);
VIR_FREE(pemdata);
return NULL;
}
#endif
static qemuMigrationCookieGraphicsPtr
qemuMigrationCookieGraphicsAlloc(virQEMUDriverPtr driver,
virDomainGraphicsDefPtr def)
{
qemuMigrationCookieGraphicsPtr mig = NULL;
const char *listenAddr;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
if (VIR_ALLOC(mig) < 0)
goto error;
mig->type = def->type;
if (mig->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC) {
mig->port = def->data.vnc.port;
listenAddr = virDomainGraphicsListenGetAddress(def, 0);
if (!listenAddr)
listenAddr = cfg->vncListen;
#ifdef WITH_GNUTLS
if (cfg->vncTLS &&
!(mig->tlsSubject = qemuDomainExtractTLSSubject(cfg->vncTLSx509certdir)))
goto error;
#endif
} else {
mig->port = def->data.spice.port;
if (cfg->spiceTLS)
mig->tlsPort = def->data.spice.tlsPort;
else
mig->tlsPort = -1;
listenAddr = virDomainGraphicsListenGetAddress(def, 0);
if (!listenAddr)
listenAddr = cfg->spiceListen;
#ifdef WITH_GNUTLS
if (cfg->spiceTLS &&
!(mig->tlsSubject = qemuDomainExtractTLSSubject(cfg->spiceTLSx509certdir)))
goto error;
#endif
}
if (VIR_STRDUP(mig->listen, listenAddr) < 0)
goto error;
virObjectUnref(cfg);
return mig;
error:
qemuMigrationCookieGraphicsFree(mig);
virObjectUnref(cfg);
return NULL;
}
static qemuMigrationCookieNetworkPtr
qemuMigrationCookieNetworkAlloc(virQEMUDriverPtr driver ATTRIBUTE_UNUSED,
virDomainDefPtr def)
{
qemuMigrationCookieNetworkPtr mig;
size_t i;
if (VIR_ALLOC(mig) < 0)
goto error;
mig->nnets = def->nnets;
if (VIR_ALLOC_N(mig->net, def->nnets) <0)
goto error;
for (i = 0; i < def->nnets; i++) {
virDomainNetDefPtr netptr;
virNetDevVPortProfilePtr vport;
netptr = def->nets[i];
vport = virDomainNetGetActualVirtPortProfile(netptr);
if (vport) {
mig->net[i].vporttype = vport->virtPortType;
switch (vport->virtPortType) {
case VIR_NETDEV_VPORT_PROFILE_NONE:
case VIR_NETDEV_VPORT_PROFILE_8021QBG:
case VIR_NETDEV_VPORT_PROFILE_8021QBH:
break;
case VIR_NETDEV_VPORT_PROFILE_OPENVSWITCH:
if (virNetDevOpenvswitchGetMigrateData(&mig->net[i].portdata,
netptr->ifname) != 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Unable to run command to get OVS port data for "
"interface %s"), netptr->ifname);
goto error;
}
break;
default:
break;
}
}
}
return mig;
error:
qemuMigrationCookieNetworkFree(mig);
return NULL;
}
static qemuMigrationCookiePtr
qemuMigrationCookieNew(virDomainObjPtr dom)
{
qemuDomainObjPrivatePtr priv = dom->privateData;
qemuMigrationCookiePtr mig = NULL;
const char *name;
if (VIR_ALLOC(mig) < 0)
goto error;
if (priv->origname)
name = priv->origname;
else
name = dom->def->name;
if (VIR_STRDUP(mig->name, name) < 0)
goto error;
memcpy(mig->uuid, dom->def->uuid, VIR_UUID_BUFLEN);
if (!(mig->localHostname = virGetHostname()))
goto error;
if (virGetHostUUID(mig->localHostuuid) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Unable to obtain host UUID"));
goto error;
}
return mig;
error:
qemuMigrationCookieFree(mig);
return NULL;
}
static int
qemuMigrationCookieAddGraphics(qemuMigrationCookiePtr mig,
virQEMUDriverPtr driver,
virDomainObjPtr dom)
{
size_t i = 0;
if (mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Migration graphics data already present"));
return -1;
}
for (i = 0; i < dom->def->ngraphics; i++) {
if (dom->def->graphics[i]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
if (!(mig->graphics =
qemuMigrationCookieGraphicsAlloc(driver, dom->def->graphics[i])))
return -1;
mig->flags |= QEMU_MIGRATION_COOKIE_GRAPHICS;
break;
}
}
return 0;
}
static int
qemuMigrationCookieAddLockstate(qemuMigrationCookiePtr mig,
virQEMUDriverPtr driver,
virDomainObjPtr dom)
{
qemuDomainObjPrivatePtr priv = dom->privateData;
if (mig->flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Migration lockstate data already present"));
return -1;
}
if (virDomainObjGetState(dom, NULL) == VIR_DOMAIN_PAUSED) {
if (VIR_STRDUP(mig->lockState, priv->lockState) < 0)
return -1;
} else {
if (virDomainLockProcessInquire(driver->lockManager, dom, &mig->lockState) < 0)
return -1;
}
if (VIR_STRDUP(mig->lockDriver, virLockManagerPluginGetName(driver->lockManager)) < 0) {
VIR_FREE(mig->lockState);
return -1;
}
mig->flags |= QEMU_MIGRATION_COOKIE_LOCKSTATE;
mig->flagsMandatory |= QEMU_MIGRATION_COOKIE_LOCKSTATE;
return 0;
}
static int
qemuMigrationCookieAddPersistent(qemuMigrationCookiePtr mig,
virDomainObjPtr dom)
{
if (mig->flags & QEMU_MIGRATION_COOKIE_PERSISTENT) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Migration persistent data already present"));
return -1;
}
if (!dom->newDef)
return 0;
mig->persistent = dom->newDef;
mig->flags |= QEMU_MIGRATION_COOKIE_PERSISTENT;
mig->flagsMandatory |= QEMU_MIGRATION_COOKIE_PERSISTENT;
return 0;
}
static virDomainDefPtr
qemuMigrationCookieGetPersistent(qemuMigrationCookiePtr mig)
{
virDomainDefPtr def = mig->persistent;
mig->persistent = NULL;
mig->flags &= ~QEMU_MIGRATION_COOKIE_PERSISTENT;
mig->flagsMandatory &= ~QEMU_MIGRATION_COOKIE_PERSISTENT;
return def;
}
static int
qemuMigrationCookieAddNetwork(qemuMigrationCookiePtr mig,
virQEMUDriverPtr driver,
virDomainObjPtr dom)
{
if (mig->flags & QEMU_MIGRATION_COOKIE_NETWORK) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Network migration data already present"));
return -1;
}
if (dom->def->nnets > 0) {
mig->network = qemuMigrationCookieNetworkAlloc(driver, dom->def);
if (!mig->network)
return -1;
mig->flags |= QEMU_MIGRATION_COOKIE_NETWORK;
}
return 0;
}
static int
qemuMigrationCookieAddNBD(qemuMigrationCookiePtr mig,
virQEMUDriverPtr driver,
virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
virHashTablePtr stats = NULL;
size_t i;
int ret = -1, rc;
/* It is not a bug if there already is a NBD data */
qemuMigrationCookieNBDFree(mig->nbd);
if (VIR_ALLOC(mig->nbd) < 0)
return -1;
if (vm->def->ndisks &&
VIR_ALLOC_N(mig->nbd->disks, vm->def->ndisks) < 0)
return -1;
mig->nbd->ndisks = 0;
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
qemuBlockStats *entry;
if (!stats) {
if (!(stats = virHashCreate(10, virHashValueFree)))
goto cleanup;
if (qemuDomainObjEnterMonitorAsync(driver, vm,
priv->job.asyncJob) < 0)
goto cleanup;
rc = qemuMonitorBlockStatsUpdateCapacity(priv->mon, stats, false);
if (qemuDomainObjExitMonitor(driver, vm) < 0)
goto cleanup;
if (rc < 0)
goto cleanup;
}
if (!disk->info.alias ||
!(entry = virHashLookup(stats, disk->info.alias)))
continue;
if (VIR_STRDUP(mig->nbd->disks[mig->nbd->ndisks].target,
disk->dst) < 0)
goto cleanup;
mig->nbd->disks[mig->nbd->ndisks].capacity = entry->capacity;
mig->nbd->ndisks++;
}
mig->nbd->port = priv->nbdPort;
mig->flags |= QEMU_MIGRATION_COOKIE_NBD;
ret = 0;
cleanup:
virHashFree(stats);
return ret;
}
static int
qemuMigrationCookieAddStatistics(qemuMigrationCookiePtr mig,
virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (!priv->job.completed)
return 0;
if (!mig->jobInfo && VIR_ALLOC(mig->jobInfo) < 0)
return -1;
*mig->jobInfo = *priv->job.completed;
mig->flags |= QEMU_MIGRATION_COOKIE_STATS;
return 0;
}
static void qemuMigrationCookieGraphicsXMLFormat(virBufferPtr buf,
qemuMigrationCookieGraphicsPtr grap)
{
virBufferAsprintf(buf, "type),
grap->port, grap->listen);
if (grap->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE)
virBufferAsprintf(buf, " tlsPort='%d'", grap->tlsPort);
if (grap->tlsSubject) {
virBufferAddLit(buf, ">\n");
virBufferAdjustIndent(buf, 2);
virBufferEscapeString(buf, "\n", grap->tlsSubject);
virBufferAdjustIndent(buf, -2);
virBufferAddLit(buf, "\n");
} else {
virBufferAddLit(buf, "/>\n");
}
}
static void
qemuMigrationCookieNetworkXMLFormat(virBufferPtr buf,
qemuMigrationCookieNetworkPtr optr)
{
size_t i;
bool empty = true;
for (i = 0; i < optr->nnets; i++) {
/* If optr->net[i].vporttype is not set, there is nothing to transfer */
if (optr->net[i].vporttype != VIR_NETDEV_VPORT_PROFILE_NONE) {
if (empty) {
virBufferAddLit(buf, "\n");
virBufferAdjustIndent(buf, 2);
empty = false;
}
virBufferAsprintf(buf, "net[i].vporttype));
if (optr->net[i].portdata) {
virBufferAddLit(buf, ">\n");
virBufferAdjustIndent(buf, 2);
virBufferEscapeString(buf, "%s\n",
optr->net[i].portdata);
virBufferAdjustIndent(buf, -2);
virBufferAddLit(buf, "\n");
} else {
virBufferAddLit(buf, "/>\n");
}
}
}
if (!empty) {
virBufferAdjustIndent(buf, -2);
virBufferAddLit(buf, "\n");
}
}
static void
qemuMigrationCookieStatisticsXMLFormat(virBufferPtr buf,
qemuDomainJobInfoPtr jobInfo)
{
qemuMonitorMigrationStats *stats = &jobInfo->stats;
virBufferAddLit(buf, "\n");
virBufferAdjustIndent(buf, 2);
virBufferAsprintf(buf, "%llu\n", jobInfo->started);
virBufferAsprintf(buf, "%llu\n", jobInfo->stopped);
virBufferAsprintf(buf, "%llu\n", jobInfo->sent);
if (jobInfo->timeDeltaSet)
virBufferAsprintf(buf, "%lld\n", jobInfo->timeDelta);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_TIME_ELAPSED,
jobInfo->timeElapsed);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_TIME_REMAINING,
jobInfo->timeRemaining);
if (stats->downtime_set)
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_DOWNTIME,
stats->downtime);
if (stats->setup_time_set)
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_SETUP_TIME,
stats->setup_time);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_TOTAL,
stats->ram_total);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_PROCESSED,
stats->ram_transferred);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_REMAINING,
stats->ram_remaining);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_BPS,
stats->ram_bps);
if (stats->ram_duplicate_set) {
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_CONSTANT,
stats->ram_duplicate);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_NORMAL,
stats->ram_normal);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_NORMAL_BYTES,
stats->ram_normal_bytes);
}
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_DIRTY_RATE,
stats->ram_dirty_rate);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_MEMORY_ITERATION,
stats->ram_iteration);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_DISK_TOTAL,
stats->disk_total);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_DISK_PROCESSED,
stats->disk_transferred);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_DISK_REMAINING,
stats->disk_remaining);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_DISK_BPS,
stats->disk_bps);
if (stats->xbzrle_set) {
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_COMPRESSION_CACHE,
stats->xbzrle_cache_size);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_COMPRESSION_BYTES,
stats->xbzrle_bytes);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_COMPRESSION_PAGES,
stats->xbzrle_pages);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_COMPRESSION_CACHE_MISSES,
stats->xbzrle_cache_miss);
virBufferAsprintf(buf, "<%1$s>%2$llu%1$s>\n",
VIR_DOMAIN_JOB_COMPRESSION_OVERFLOW,
stats->xbzrle_overflow);
}
virBufferAdjustIndent(buf, -2);
virBufferAddLit(buf, "\n");
}
static int
qemuMigrationCookieXMLFormat(virQEMUDriverPtr driver,
virBufferPtr buf,
qemuMigrationCookiePtr mig)
{
char uuidstr[VIR_UUID_STRING_BUFLEN];
char hostuuidstr[VIR_UUID_STRING_BUFLEN];
size_t i;
virUUIDFormat(mig->uuid, uuidstr);
virUUIDFormat(mig->localHostuuid, hostuuidstr);
virBufferAddLit(buf, "\n");
virBufferAdjustIndent(buf, 2);
virBufferEscapeString(buf, "%s\n", mig->name);
virBufferAsprintf(buf, "%s\n", uuidstr);
virBufferEscapeString(buf, "%s\n", mig->localHostname);
virBufferAsprintf(buf, "%s\n", hostuuidstr);
for (i = 0; i < QEMU_MIGRATION_COOKIE_FLAG_LAST; i++) {
if (mig->flagsMandatory & (1 << i))
virBufferAsprintf(buf, "\n",
qemuMigrationCookieFlagTypeToString(i));
}
if ((mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) &&
mig->graphics)
qemuMigrationCookieGraphicsXMLFormat(buf, mig->graphics);
if ((mig->flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) &&
mig->lockState) {
virBufferAsprintf(buf, "\n",
mig->lockDriver);
virBufferAdjustIndent(buf, 2);
virBufferAsprintf(buf, "%s\n",
mig->lockState);
virBufferAdjustIndent(buf, -2);
virBufferAddLit(buf, "\n");
}
if ((mig->flags & QEMU_MIGRATION_COOKIE_PERSISTENT) &&
mig->persistent) {
if (qemuDomainDefFormatBuf(driver,
mig->persistent,
VIR_DOMAIN_XML_INACTIVE |
VIR_DOMAIN_XML_SECURE |
VIR_DOMAIN_XML_MIGRATABLE,
buf) < 0)
return -1;
}
if ((mig->flags & QEMU_MIGRATION_COOKIE_NETWORK) && mig->network)
qemuMigrationCookieNetworkXMLFormat(buf, mig->network);
if ((mig->flags & QEMU_MIGRATION_COOKIE_NBD) && mig->nbd) {
virBufferAddLit(buf, "nbd->port)
virBufferAsprintf(buf, " port='%d'", mig->nbd->port);
if (mig->nbd->ndisks) {
virBufferAddLit(buf, ">\n");
virBufferAdjustIndent(buf, 2);
for (i = 0; i < mig->nbd->ndisks; i++) {
virBufferEscapeString(buf, "nbd->disks[i].target);
virBufferAsprintf(buf, " capacity='%llu'/>\n",
mig->nbd->disks[i].capacity);
}
virBufferAdjustIndent(buf, -2);
virBufferAddLit(buf, "\n");
} else {
virBufferAddLit(buf, "/>\n");
}
}
if (mig->flags & QEMU_MIGRATION_COOKIE_STATS && mig->jobInfo)
qemuMigrationCookieStatisticsXMLFormat(buf, mig->jobInfo);
virBufferAdjustIndent(buf, -2);
virBufferAddLit(buf, "\n");
return 0;
}
static char *qemuMigrationCookieXMLFormatStr(virQEMUDriverPtr driver,
qemuMigrationCookiePtr mig)
{
virBuffer buf = VIR_BUFFER_INITIALIZER;
if (qemuMigrationCookieXMLFormat(driver, &buf, mig) < 0) {
virBufferFreeAndReset(&buf);
return NULL;
}
if (virBufferCheckError(&buf) < 0)
return NULL;
return virBufferContentAndReset(&buf);
}
static qemuMigrationCookieGraphicsPtr
qemuMigrationCookieGraphicsXMLParse(xmlXPathContextPtr ctxt)
{
qemuMigrationCookieGraphicsPtr grap;
char *tmp;
if (VIR_ALLOC(grap) < 0)
goto error;
if (!(tmp = virXPathString("string(./graphics/@type)", ctxt))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing type attribute in migration data"));
goto error;
}
if ((grap->type = virDomainGraphicsTypeFromString(tmp)) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unknown graphics type %s"), tmp);
VIR_FREE(tmp);
goto error;
}
VIR_FREE(tmp);
if (virXPathInt("string(./graphics/@port)", ctxt, &grap->port) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing port attribute in migration data"));
goto error;
}
if (grap->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
if (virXPathInt("string(./graphics/@tlsPort)", ctxt, &grap->tlsPort) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing tlsPort attribute in migration data"));
goto error;
}
}
if (!(grap->listen = virXPathString("string(./graphics/@listen)", ctxt))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing listen attribute in migration data"));
goto error;
}
/* Optional */
grap->tlsSubject = virXPathString("string(./graphics/cert[@info='subject']/@value)", ctxt);
return grap;
error:
qemuMigrationCookieGraphicsFree(grap);
return NULL;
}
static qemuMigrationCookieNetworkPtr
qemuMigrationCookieNetworkXMLParse(xmlXPathContextPtr ctxt)
{
qemuMigrationCookieNetworkPtr optr;
size_t i;
int n;
xmlNodePtr *interfaces = NULL;
char *vporttype;
xmlNodePtr save_ctxt = ctxt->node;
if (VIR_ALLOC(optr) < 0)
goto error;
if ((n = virXPathNodeSet("./network/interface", ctxt, &interfaces)) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing interface information"));
goto error;
}
optr->nnets = n;
if (VIR_ALLOC_N(optr->net, optr->nnets) < 0)
goto error;
for (i = 0; i < n; i++) {
/* portdata is optional, and may not exist */
ctxt->node = interfaces[i];
optr->net[i].portdata = virXPathString("string(./portdata[1])", ctxt);
if (!(vporttype = virXMLPropString(interfaces[i], "vporttype"))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing vporttype attribute in migration data"));
goto error;
}
optr->net[i].vporttype = virNetDevVPortTypeFromString(vporttype);
}
VIR_FREE(interfaces);
cleanup:
ctxt->node = save_ctxt;
return optr;
error:
VIR_FREE(interfaces);
qemuMigrationCookieNetworkFree(optr);
optr = NULL;
goto cleanup;
}
static qemuMigrationCookieNBDPtr
qemuMigrationCookieNBDXMLParse(xmlXPathContextPtr ctxt)
{
qemuMigrationCookieNBDPtr ret = NULL;
char *port = NULL, *capacity = NULL;
size_t i;
int n;
xmlNodePtr *disks = NULL;
xmlNodePtr save_ctxt = ctxt->node;
if (VIR_ALLOC(ret) < 0)
goto error;
port = virXPathString("string(./nbd/@port)", ctxt);
if (port && virStrToLong_i(port, NULL, 10, &ret->port) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Malformed nbd port '%s'"),
port);
goto error;
}
/* Now check if source sent a list of disks to prealloc. We might be
* talking to an older server, so it's not an error if the list is
* missing. */
if ((n = virXPathNodeSet("./nbd/disk", ctxt, &disks)) > 0) {
if (VIR_ALLOC_N(ret->disks, n) < 0)
goto error;
ret->ndisks = n;
for (i = 0; i < n; i++) {
ctxt->node = disks[i];
VIR_FREE(capacity);
if (!(ret->disks[i].target = virXPathString("string(./@target)", ctxt))) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Malformed disk target"));
goto error;
}
capacity = virXPathString("string(./@capacity)", ctxt);
if (!capacity ||
virStrToLong_ull(capacity, NULL, 10,
&ret->disks[i].capacity) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Malformed disk capacity: '%s'"),
NULLSTR(capacity));
goto error;
}
}
}
cleanup:
VIR_FREE(port);
VIR_FREE(capacity);
VIR_FREE(disks);
ctxt->node = save_ctxt;
return ret;
error:
qemuMigrationCookieNBDFree(ret);
ret = NULL;
goto cleanup;
}
static qemuDomainJobInfoPtr
qemuMigrationCookieStatisticsXMLParse(xmlXPathContextPtr ctxt)
{
qemuDomainJobInfoPtr jobInfo = NULL;
qemuMonitorMigrationStats *stats;
xmlNodePtr save_ctxt = ctxt->node;
if (!(ctxt->node = virXPathNode("./statistics", ctxt)))
goto cleanup;
if (VIR_ALLOC(jobInfo) < 0)
goto cleanup;
stats = &jobInfo->stats;
jobInfo->type = VIR_DOMAIN_JOB_COMPLETED;
virXPathULongLong("string(./started[1])", ctxt, &jobInfo->started);
virXPathULongLong("string(./stopped[1])", ctxt, &jobInfo->stopped);
virXPathULongLong("string(./sent[1])", ctxt, &jobInfo->sent);
if (virXPathLongLong("string(./delta[1])", ctxt, &jobInfo->timeDelta) == 0)
jobInfo->timeDeltaSet = true;
virXPathULongLong("string(./" VIR_DOMAIN_JOB_TIME_ELAPSED "[1])",
ctxt, &jobInfo->timeElapsed);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_TIME_REMAINING "[1])",
ctxt, &jobInfo->timeRemaining);
if (virXPathULongLong("string(./" VIR_DOMAIN_JOB_DOWNTIME "[1])",
ctxt, &stats->downtime) == 0)
stats->downtime_set = true;
if (virXPathULongLong("string(./" VIR_DOMAIN_JOB_SETUP_TIME "[1])",
ctxt, &stats->setup_time) == 0)
stats->setup_time_set = true;
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_TOTAL "[1])",
ctxt, &stats->ram_total);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_PROCESSED "[1])",
ctxt, &stats->ram_transferred);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_REMAINING "[1])",
ctxt, &stats->ram_remaining);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_BPS "[1])",
ctxt, &stats->ram_bps);
if (virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_CONSTANT "[1])",
ctxt, &stats->ram_duplicate) == 0)
stats->ram_duplicate_set = true;
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_NORMAL "[1])",
ctxt, &stats->ram_normal);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_NORMAL_BYTES "[1])",
ctxt, &stats->ram_normal_bytes);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_DIRTY_RATE "[1])",
ctxt, &stats->ram_dirty_rate);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_MEMORY_ITERATION "[1])",
ctxt, &stats->ram_iteration);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_DISK_TOTAL "[1])",
ctxt, &stats->disk_total);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_DISK_PROCESSED "[1])",
ctxt, &stats->disk_transferred);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_DISK_REMAINING "[1])",
ctxt, &stats->disk_remaining);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_DISK_BPS "[1])",
ctxt, &stats->disk_bps);
if (virXPathULongLong("string(./" VIR_DOMAIN_JOB_COMPRESSION_CACHE "[1])",
ctxt, &stats->xbzrle_cache_size) == 0)
stats->xbzrle_set = true;
virXPathULongLong("string(./" VIR_DOMAIN_JOB_COMPRESSION_BYTES "[1])",
ctxt, &stats->xbzrle_bytes);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_COMPRESSION_PAGES "[1])",
ctxt, &stats->xbzrle_pages);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_COMPRESSION_CACHE_MISSES "[1])",
ctxt, &stats->xbzrle_cache_miss);
virXPathULongLong("string(./" VIR_DOMAIN_JOB_COMPRESSION_OVERFLOW "[1])",
ctxt, &stats->xbzrle_overflow);
cleanup:
ctxt->node = save_ctxt;
return jobInfo;
}
static int
qemuMigrationCookieXMLParse(qemuMigrationCookiePtr mig,
virQEMUDriverPtr driver,
xmlDocPtr doc,
xmlXPathContextPtr ctxt,
unsigned int flags)
{
char uuidstr[VIR_UUID_STRING_BUFLEN];
char *tmp = NULL;
xmlNodePtr *nodes = NULL;
size_t i;
int n;
virCapsPtr caps = NULL;
if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
goto error;
/* We don't store the uuid, name, hostname, or hostuuid
* values. We just compare them to local data to do some
* sanity checking on migration operation
*/
/* Extract domain name */
if (!(tmp = virXPathString("string(./name[1])", ctxt))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing name element in migration data"));
goto error;
}
if (STRNEQ(tmp, mig->name)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Incoming cookie data had unexpected name %s vs %s"),
tmp, mig->name);
goto error;
}
VIR_FREE(tmp);
/* Extract domain uuid */
tmp = virXPathString("string(./uuid[1])", ctxt);
if (!tmp) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing uuid element in migration data"));
goto error;
}
virUUIDFormat(mig->uuid, uuidstr);
if (STRNEQ(tmp, uuidstr)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Incoming cookie data had unexpected UUID %s vs %s"),
tmp, uuidstr);
goto error;
}
VIR_FREE(tmp);
/* Check & forbid "localhost" migration */
if (!(mig->remoteHostname = virXPathString("string(./hostname[1])", ctxt))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing hostname element in migration data"));
goto error;
}
if (STREQ(mig->remoteHostname, mig->localHostname)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Attempt to migrate guest to the same host %s"),
mig->remoteHostname);
goto error;
}
if (!(tmp = virXPathString("string(./hostuuid[1])", ctxt))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing hostuuid element in migration data"));
goto error;
}
if (virUUIDParse(tmp, mig->remoteHostuuid) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("malformed hostuuid element in migration data"));
goto error;
}
if (memcmp(mig->remoteHostuuid, mig->localHostuuid, VIR_UUID_BUFLEN) == 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Attempt to migrate guest to the same host %s"),
tmp);
goto error;
}
VIR_FREE(tmp);
/* Check to ensure all mandatory features from XML are also
* present in 'flags' */
if ((n = virXPathNodeSet("./feature", ctxt, &nodes)) < 0)
goto error;
for (i = 0; i < n; i++) {
int val;
char *str = virXMLPropString(nodes[i], "name");
if (!str) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("missing feature name"));
goto error;
}
if ((val = qemuMigrationCookieFlagTypeFromString(str)) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Unknown migration cookie feature %s"),
str);
VIR_FREE(str);
goto error;
}
if ((flags & (1 << val)) == 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Unsupported migration cookie feature %s"),
str);
VIR_FREE(str);
goto error;
}
VIR_FREE(str);
}
VIR_FREE(nodes);
if ((flags & QEMU_MIGRATION_COOKIE_GRAPHICS) &&
virXPathBoolean("count(./graphics) > 0", ctxt) &&
(!(mig->graphics = qemuMigrationCookieGraphicsXMLParse(ctxt))))
goto error;
if ((flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) &&
virXPathBoolean("count(./lockstate) > 0", ctxt)) {
mig->lockDriver = virXPathString("string(./lockstate[1]/@driver)", ctxt);
if (!mig->lockDriver) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Missing lock driver name in migration cookie"));
goto error;
}
mig->lockState = virXPathString("string(./lockstate[1]/leases[1])", ctxt);
if (mig->lockState && STREQ(mig->lockState, ""))
VIR_FREE(mig->lockState);
}
if ((flags & QEMU_MIGRATION_COOKIE_PERSISTENT) &&
virXPathBoolean("count(./domain) > 0", ctxt)) {
if ((n = virXPathNodeSet("./domain", ctxt, &nodes)) > 1) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Too many domain elements in "
"migration cookie: %d"),
n);
goto error;
}
mig->persistent = virDomainDefParseNode(doc, nodes[0],
caps, driver->xmlopt,
VIR_DOMAIN_DEF_PARSE_INACTIVE |
VIR_DOMAIN_DEF_PARSE_ABI_UPDATE);
if (!mig->persistent) {
/* virDomainDefParseNode already reported
* an error for us */
goto error;
}
VIR_FREE(nodes);
}
if ((flags & QEMU_MIGRATION_COOKIE_NETWORK) &&
virXPathBoolean("count(./network) > 0", ctxt) &&
(!(mig->network = qemuMigrationCookieNetworkXMLParse(ctxt))))
goto error;
if (flags & QEMU_MIGRATION_COOKIE_NBD &&
virXPathBoolean("boolean(./nbd)", ctxt) &&
(!(mig->nbd = qemuMigrationCookieNBDXMLParse(ctxt))))
goto error;
if (flags & QEMU_MIGRATION_COOKIE_STATS &&
virXPathBoolean("boolean(./statistics)", ctxt) &&
(!(mig->jobInfo = qemuMigrationCookieStatisticsXMLParse(ctxt))))
goto error;
virObjectUnref(caps);
return 0;
error:
VIR_FREE(tmp);
VIR_FREE(nodes);
virObjectUnref(caps);
return -1;
}
static int
qemuMigrationCookieXMLParseStr(qemuMigrationCookiePtr mig,
virQEMUDriverPtr driver,
const char *xml,
unsigned int flags)
{
xmlDocPtr doc = NULL;
xmlXPathContextPtr ctxt = NULL;
int ret = -1;
VIR_DEBUG("xml=%s", NULLSTR(xml));
if (!(doc = virXMLParseStringCtxt(xml, _("(qemu_migration_cookie)"), &ctxt)))
goto cleanup;
ret = qemuMigrationCookieXMLParse(mig, driver, doc, ctxt, flags);
cleanup:
xmlXPathFreeContext(ctxt);
xmlFreeDoc(doc);
return ret;
}
static int
qemuMigrationBakeCookie(qemuMigrationCookiePtr mig,
virQEMUDriverPtr driver,
virDomainObjPtr dom,
char **cookieout,
int *cookieoutlen,
unsigned int flags)
{
if (!cookieout || !cookieoutlen)
return 0;
*cookieoutlen = 0;
if (flags & QEMU_MIGRATION_COOKIE_GRAPHICS &&
qemuMigrationCookieAddGraphics(mig, driver, dom) < 0)
return -1;
if (flags & QEMU_MIGRATION_COOKIE_LOCKSTATE &&
qemuMigrationCookieAddLockstate(mig, driver, dom) < 0)
return -1;
if (flags & QEMU_MIGRATION_COOKIE_PERSISTENT &&
qemuMigrationCookieAddPersistent(mig, dom) < 0)
return -1;
if (flags & QEMU_MIGRATION_COOKIE_NETWORK &&
qemuMigrationCookieAddNetwork(mig, driver, dom) < 0) {
return -1;
}
if ((flags & QEMU_MIGRATION_COOKIE_NBD) &&
qemuMigrationCookieAddNBD(mig, driver, dom) < 0)
return -1;
if (flags & QEMU_MIGRATION_COOKIE_STATS &&
qemuMigrationCookieAddStatistics(mig, dom) < 0)
return -1;
if (flags & QEMU_MIGRATION_COOKIE_MEMORY_HOTPLUG)
mig->flagsMandatory |= QEMU_MIGRATION_COOKIE_MEMORY_HOTPLUG;
if (!(*cookieout = qemuMigrationCookieXMLFormatStr(driver, mig)))
return -1;
*cookieoutlen = strlen(*cookieout) + 1;
VIR_DEBUG("cookielen=%d cookie=%s", *cookieoutlen, *cookieout);
return 0;
}
static qemuMigrationCookiePtr
qemuMigrationEatCookie(virQEMUDriverPtr driver,
virDomainObjPtr dom,
const char *cookiein,
int cookieinlen,
unsigned int flags)
{
qemuMigrationCookiePtr mig = NULL;
/* Parse & validate incoming cookie (if any) */
if (cookiein && cookieinlen &&
cookiein[cookieinlen-1] != '\0') {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Migration cookie was not NULL terminated"));
goto error;
}
VIR_DEBUG("cookielen=%d cookie='%s'", cookieinlen, NULLSTR(cookiein));
if (!(mig = qemuMigrationCookieNew(dom)))
return NULL;
if (cookiein && cookieinlen &&
qemuMigrationCookieXMLParseStr(mig,
driver,
cookiein,
flags) < 0)
goto error;
if (flags & QEMU_MIGRATION_COOKIE_PERSISTENT &&
mig->persistent &&
STRNEQ(dom->def->name, mig->persistent->name)) {
VIR_FREE(mig->persistent->name);
if (VIR_STRDUP(mig->persistent->name, dom->def->name) < 0)
goto error;
}
if (mig->flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) {
if (!mig->lockDriver) {
if (virLockManagerPluginUsesState(driver->lockManager)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Missing %s lock state for migration cookie"),
virLockManagerPluginGetName(driver->lockManager));
goto error;
}
} else if (STRNEQ(mig->lockDriver,
virLockManagerPluginGetName(driver->lockManager))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Source host lock driver %s different from target %s"),
mig->lockDriver,
virLockManagerPluginGetName(driver->lockManager));
goto error;
}
}
return mig;
error:
qemuMigrationCookieFree(mig);
return NULL;
}
static void
qemuMigrationStoreDomainState(virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
priv->preMigrationState = virDomainObjGetState(vm, NULL);
VIR_DEBUG("Storing pre-migration state=%d domain=%p",
priv->preMigrationState, vm);
}
/* Returns true if the domain was resumed, false otherwise */
static bool
qemuMigrationRestoreDomainState(virConnectPtr conn, virDomainObjPtr vm)
{
virQEMUDriverPtr driver = conn->privateData;
qemuDomainObjPrivatePtr priv = vm->privateData;
int reason;
virDomainState state = virDomainObjGetState(vm, &reason);
bool ret = false;
VIR_DEBUG("driver=%p, vm=%p, pre-mig-state=%s, state=%s, reason=%s",
driver, vm,
virDomainStateTypeToString(priv->preMigrationState),
virDomainStateTypeToString(state),
virDomainStateReasonToString(state, reason));
if (state != VIR_DOMAIN_PAUSED ||
reason == VIR_DOMAIN_PAUSED_POSTCOPY_FAILED)
goto cleanup;
if (priv->preMigrationState == VIR_DOMAIN_RUNNING) {
/* This is basically the only restore possibility that's safe
* and we should attempt to do */
VIR_DEBUG("Restoring pre-migration state due to migration error");
/* we got here through some sort of failure; start the domain again */
if (qemuProcessStartCPUs(driver, vm, conn,
VIR_DOMAIN_RUNNING_MIGRATION_CANCELED,
QEMU_ASYNC_JOB_MIGRATION_OUT) < 0) {
/* Hm, we already know we are in error here. We don't want to
* overwrite the previous error, though, so we just throw something
* to the logs and hope for the best */
VIR_ERROR(_("Failed to resume guest %s after failure"), vm->def->name);
goto cleanup;
}
ret = true;
}
cleanup:
priv->preMigrationState = VIR_DOMAIN_NOSTATE;
return ret;
}
static int
qemuMigrationPrecreateDisk(virConnectPtr conn,
virDomainDiskDefPtr disk,
unsigned long long capacity)
{
int ret = -1;
virStoragePoolPtr pool = NULL;
virStorageVolPtr vol = NULL;
char *volName = NULL, *basePath = NULL;
char *volStr = NULL;
virBuffer buf = VIR_BUFFER_INITIALIZER;
const char *format = NULL;
unsigned int flags = 0;
VIR_DEBUG("Precreate disk type=%s", virStorageTypeToString(disk->src->type));
switch ((virStorageType) disk->src->type) {
case VIR_STORAGE_TYPE_FILE:
if (!virDomainDiskGetSource(disk)) {
VIR_DEBUG("Dropping sourceless disk '%s'",
disk->dst);
return 0;
}
if (VIR_STRDUP(basePath, disk->src->path) < 0)
goto cleanup;
if (!(volName = strrchr(basePath, '/'))) {
virReportError(VIR_ERR_INVALID_ARG,
_("malformed disk path: %s"),
disk->src->path);
goto cleanup;
}
*volName = '\0';
volName++;
if (!(pool = storagePoolLookupByTargetPath(conn, basePath)))
goto cleanup;
format = virStorageFileFormatTypeToString(disk->src->format);
if (disk->src->format == VIR_STORAGE_FILE_QCOW2)
flags |= VIR_STORAGE_VOL_CREATE_PREALLOC_METADATA;
break;
case VIR_STORAGE_TYPE_VOLUME:
if (!(pool = virStoragePoolLookupByName(conn, disk->src->srcpool->pool)))
goto cleanup;
format = virStorageFileFormatTypeToString(disk->src->format);
volName = disk->src->srcpool->volume;
if (disk->src->format == VIR_STORAGE_FILE_QCOW2)
flags |= VIR_STORAGE_VOL_CREATE_PREALLOC_METADATA;
break;
case VIR_STORAGE_TYPE_NETWORK:
VIR_DEBUG("Skipping creation of network disk '%s'",
disk->dst);
return 0;
case VIR_STORAGE_TYPE_BLOCK:
case VIR_STORAGE_TYPE_DIR:
case VIR_STORAGE_TYPE_NONE:
case VIR_STORAGE_TYPE_LAST:
virReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot precreate storage for disk type '%s'"),
virStorageTypeToString(disk->src->type));
goto cleanup;
break;
}
if ((vol = virStorageVolLookupByName(pool, volName))) {
VIR_DEBUG("Skipping creation of already existing volume of name '%s'",
volName);
ret = 0;
goto cleanup;
}
virBufferAddLit(&buf, "\n");
virBufferAdjustIndent(&buf, 2);
virBufferEscapeString(&buf, "%s\n", volName);
virBufferAsprintf(&buf, "%llu\n", capacity);
virBufferAddLit(&buf, "\n");
virBufferAdjustIndent(&buf, 2);
virBufferAsprintf(&buf, "\n", format);
virBufferAdjustIndent(&buf, -2);
virBufferAddLit(&buf, "\n");
virBufferAdjustIndent(&buf, -2);
virBufferAddLit(&buf, "\n");
if (!(volStr = virBufferContentAndReset(&buf))) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("unable to create volume XML"));
goto cleanup;
}
if (!(vol = virStorageVolCreateXML(pool, volStr, flags)))
goto cleanup;
ret = 0;
cleanup:
VIR_FREE(basePath);
VIR_FREE(volStr);
virObjectUnref(vol);
virObjectUnref(pool);
return ret;
}
static bool
qemuMigrateDisk(virDomainDiskDef const *disk,
size_t nmigrate_disks, const char **migrate_disks)
{
size_t i;
/* Check if the disk alias is in the list */
if (nmigrate_disks) {
for (i = 0; i < nmigrate_disks; i++) {
if (STREQ(disk->dst, migrate_disks[i]))
return true;
}
return false;
}
/* Default is to migrate only non-shared non-readonly disks
* with source */
return !disk->src->shared && !disk->src->readonly &&
virDomainDiskGetSource(disk);
}
static int
qemuMigrationPrecreateStorage(virConnectPtr conn,
virQEMUDriverPtr driver ATTRIBUTE_UNUSED,
virDomainObjPtr vm,
qemuMigrationCookieNBDPtr nbd,
size_t nmigrate_disks,
const char **migrate_disks,
bool incremental)
{
int ret = -1;
size_t i = 0;
if (!nbd || !nbd->ndisks)
return 0;
for (i = 0; i < nbd->ndisks; i++) {
virDomainDiskDefPtr disk;
const char *diskSrcPath;
VIR_DEBUG("Looking up disk target '%s' (capacity=%llu)",
nbd->disks[i].target, nbd->disks[i].capacity);
if (!(disk = virDomainDiskByName(vm->def, nbd->disks[i].target,
false))) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unable to find disk by target: %s"),
nbd->disks[i].target);
goto cleanup;
}
diskSrcPath = virDomainDiskGetSource(disk);
/* Skip disks we don't want to migrate and already existing disks. */
if (!qemuMigrateDisk(disk, nmigrate_disks, migrate_disks) ||
(diskSrcPath && virFileExists(diskSrcPath))) {
continue;
}
if (incremental) {
virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s",
_("pre-creation of storage targets for incremental "
"storage migration is not supported"));
goto cleanup;
}
VIR_DEBUG("Proceeding with disk source %s", NULLSTR(diskSrcPath));
if (qemuMigrationPrecreateDisk(conn, disk, nbd->disks[i].capacity) < 0)
goto cleanup;
}
ret = 0;
cleanup:
return ret;
}
/**
* qemuMigrationStartNBDServer:
* @driver: qemu driver
* @vm: domain
*
* Starts NBD server. This is a newer method to copy
* storage during migration than using 'blk' and 'inc'
* arguments in 'migrate' monitor command.
* Error is reported here.
*
* Returns 0 on success, -1 otherwise.
*/
static int
qemuMigrationStartNBDServer(virQEMUDriverPtr driver,
virDomainObjPtr vm,
const char *listenAddr,
size_t nmigrate_disks,
const char **migrate_disks,
int nbdPort)
{
int ret = -1;
qemuDomainObjPrivatePtr priv = vm->privateData;
unsigned short port = 0;
char *diskAlias = NULL;
size_t i;
if (nbdPort < 0 || nbdPort > USHRT_MAX) {
virReportError(VIR_ERR_INVALID_ARG, "%s",
_("nbd port must be in range 0-65535"));
return -1;
}
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
/* check whether disk should be migrated */
if (!qemuMigrateDisk(disk, nmigrate_disks, migrate_disks))
continue;
VIR_FREE(diskAlias);
if (virAsprintf(&diskAlias, "%s%s",
QEMU_DRIVE_HOST_PREFIX, disk->info.alias) < 0)
goto cleanup;
if (qemuDomainObjEnterMonitorAsync(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
goto cleanup;
if (port == 0) {
if (nbdPort)
port = nbdPort;
else if (virPortAllocatorAcquire(driver->migrationPorts, &port) < 0)
goto exit_monitor;
if (qemuMonitorNBDServerStart(priv->mon, listenAddr, port) < 0)
goto exit_monitor;
}
if (qemuMonitorNBDServerAdd(priv->mon, diskAlias, true) < 0)
goto exit_monitor;
if (qemuDomainObjExitMonitor(driver, vm) < 0)
goto cleanup;
}
priv->nbdPort = port;
ret = 0;
cleanup:
VIR_FREE(diskAlias);
if (ret < 0 && nbdPort == 0)
virPortAllocatorRelease(driver->migrationPorts, port);
return ret;
exit_monitor:
ignore_value(qemuDomainObjExitMonitor(driver, vm));
goto cleanup;
}
static int
qemuMigrationStopNBDServer(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMigrationCookiePtr mig)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (!mig->nbd)
return 0;
if (qemuDomainObjEnterMonitorAsync(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
return -1;
if (qemuMonitorNBDServerStop(priv->mon) < 0)
VIR_WARN("Unable to stop NBD server");
if (qemuDomainObjExitMonitor(driver, vm) < 0)
return -1;
virPortAllocatorRelease(driver->migrationPorts, priv->nbdPort);
priv->nbdPort = 0;
return 0;
}
/**
* qemuMigrationDriveMirrorReady:
* @driver: qemu driver
* @vm: domain
*
* Check the status of all drive-mirrors started by
* qemuMigrationDriveMirror. Any pending block job events
* for the mirrored disks will be processed.
*
* Returns 1 if all mirrors are "ready",
* 0 if some mirrors are still performing initial sync,
* -1 on error.
*/
static int
qemuMigrationDriveMirrorReady(virQEMUDriverPtr driver,
virDomainObjPtr vm)
{
size_t i;
size_t notReady = 0;
int status;
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
if (!diskPriv->migrating)
continue;
status = qemuBlockJobUpdate(driver, vm, disk);
if (status == VIR_DOMAIN_BLOCK_JOB_FAILED) {
virReportError(VIR_ERR_OPERATION_FAILED,
_("migration of disk %s failed"),
disk->dst);
return -1;
}
if (disk->mirrorState != VIR_DOMAIN_DISK_MIRROR_STATE_READY)
notReady++;
}
if (notReady) {
VIR_DEBUG("Waiting for %zu disk mirrors to get ready", notReady);
return 0;
} else {
VIR_DEBUG("All disk mirrors are ready");
return 1;
}
}
/*
* If @check is true, the function will report an error and return a different
* code in case a block job fails. This way we can properly abort migration in
* case some block jobs failed once all memory has already been transferred.
*
* Returns 1 if all mirrors are gone,
* 0 if some mirrors are still active,
* -1 some mirrors failed but some are still active,
* -2 all mirrors are gone but some of them failed.
*/
static int
qemuMigrationDriveMirrorCancelled(virQEMUDriverPtr driver,
virDomainObjPtr vm,
bool check)
{
size_t i;
size_t active = 0;
int status;
bool failed = false;
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
if (!diskPriv->migrating)
continue;
status = qemuBlockJobUpdate(driver, vm, disk);
switch (status) {
case VIR_DOMAIN_BLOCK_JOB_FAILED:
if (check) {
virReportError(VIR_ERR_OPERATION_FAILED,
_("migration of disk %s failed"),
disk->dst);
failed = true;
}
/* fallthrough */
case VIR_DOMAIN_BLOCK_JOB_CANCELED:
case VIR_DOMAIN_BLOCK_JOB_COMPLETED:
qemuBlockJobSyncEnd(driver, vm, disk);
diskPriv->migrating = false;
break;
default:
active++;
}
}
if (failed) {
if (active) {
VIR_DEBUG("Some disk mirrors failed; still waiting for %zu "
"disk mirrors to finish", active);
return -1;
} else {
VIR_DEBUG("All disk mirrors are gone; some of them failed");
return -2;
}
} else {
if (active) {
VIR_DEBUG("Waiting for %zu disk mirrors to finish", active);
return 0;
} else {
VIR_DEBUG("All disk mirrors are gone");
return 1;
}
}
}
/*
* Returns 0 on success,
* 1 when job is already completed or it failed and failNoJob is false,
* -1 on error or when job failed and failNoJob is true.
*/
static int
qemuMigrationCancelOneDriveMirror(virQEMUDriverPtr driver,
virDomainObjPtr vm,
virDomainDiskDefPtr disk,
bool failNoJob,
qemuDomainAsyncJob asyncJob)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
char *diskAlias = NULL;
int ret = -1;
int status;
int rv;
status = qemuBlockJobUpdate(driver, vm, disk);
switch (status) {
case VIR_DOMAIN_BLOCK_JOB_FAILED:
case VIR_DOMAIN_BLOCK_JOB_CANCELED:
if (failNoJob) {
virReportError(VIR_ERR_OPERATION_FAILED,
_("migration of disk %s failed"),
disk->dst);
return -1;
}
return 1;
case VIR_DOMAIN_BLOCK_JOB_COMPLETED:
return 1;
}
if (virAsprintf(&diskAlias, "%s%s",
QEMU_DRIVE_HOST_PREFIX, disk->info.alias) < 0)
return -1;
if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
goto cleanup;
rv = qemuMonitorBlockJobCancel(priv->mon, diskAlias, true);
if (qemuDomainObjExitMonitor(driver, vm) < 0 || rv < 0)
goto cleanup;
ret = 0;
cleanup:
VIR_FREE(diskAlias);
return ret;
}
/**
* qemuMigrationCancelDriveMirror:
* @driver: qemu driver
* @vm: domain
* @check: if true report an error when some of the mirrors fails
*
* Cancel all drive-mirrors started by qemuMigrationDriveMirror.
* Any pending block job events for the affected disks will be
* processed.
*
* Returns 0 on success, -1 otherwise.
*/
static int
qemuMigrationCancelDriveMirror(virQEMUDriverPtr driver,
virDomainObjPtr vm,
bool check,
qemuDomainAsyncJob asyncJob,
virConnectPtr dconn)
{
virErrorPtr err = NULL;
int ret = -1;
size_t i;
int rv;
bool failed = false;
VIR_DEBUG("Cancelling drive mirrors for domain %s", vm->def->name);
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
if (!diskPriv->migrating)
continue;
rv = qemuMigrationCancelOneDriveMirror(driver, vm, disk,
check, asyncJob);
if (rv != 0) {
if (rv < 0) {
if (!err)
err = virSaveLastError();
failed = true;
}
qemuBlockJobSyncEnd(driver, vm, disk);
diskPriv->migrating = false;
}
}
while ((rv = qemuMigrationDriveMirrorCancelled(driver, vm, check)) != 1) {
if (check && !failed &&
dconn && virConnectIsAlive(dconn) <= 0) {
virReportError(VIR_ERR_OPERATION_FAILED, "%s",
_("Lost connection to destination host"));
failed = true;
}
if (rv < 0) {
failed = true;
if (rv == -2)
break;
}
if (failed && !err)
err = virSaveLastError();
if (virDomainObjWait(vm) < 0)
goto cleanup;
}
ret = failed ? -1 : 0;
cleanup:
if (err) {
virSetError(err);
virFreeError(err);
}
return ret;
}
/**
* qemuMigrationDriveMirror:
* @driver: qemu driver
* @vm: domain
* @mig: migration cookie
* @host: where are we migrating to
* @speed: how much should the copying be limited
* @migrate_flags: migrate monitor command flags
*
* Run drive-mirror to feed NBD server running on dst and wait
* till the process switches into another phase where writes go
* simultaneously to both source and destination. On success,
* update @migrate_flags so we don't tell 'migrate' command
* to do the very same operation. On failure, the caller is
* expected to call qemuMigrationCancelDriveMirror to stop all
* running mirrors.
*
* Returns 0 on success (@migrate_flags updated),
* -1 otherwise.
*/
static int
qemuMigrationDriveMirror(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMigrationCookiePtr mig,
const char *host,
unsigned long speed,
unsigned int *migrate_flags,
size_t nmigrate_disks,
const char **migrate_disks,
virConnectPtr dconn)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int ret = -1;
int port;
size_t i;
char *diskAlias = NULL;
char *nbd_dest = NULL;
char *hoststr = NULL;
unsigned int mirror_flags = VIR_DOMAIN_BLOCK_REBASE_REUSE_EXT;
int rv;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
VIR_DEBUG("Starting drive mirrors for domain %s", vm->def->name);
/* steal NBD port and thus prevent its propagation back to destination */
port = mig->nbd->port;
mig->nbd->port = 0;
/* escape literal IPv6 address */
if (strchr(host, ':')) {
if (virAsprintf(&hoststr, "[%s]", host) < 0)
goto cleanup;
} else if (VIR_STRDUP(hoststr, host) < 0) {
goto cleanup;
}
if (*migrate_flags & QEMU_MONITOR_MIGRATE_NON_SHARED_INC)
mirror_flags |= VIR_DOMAIN_BLOCK_REBASE_SHALLOW;
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
int mon_ret;
/* check whether disk should be migrated */
if (!qemuMigrateDisk(disk, nmigrate_disks, migrate_disks))
continue;
if ((virAsprintf(&diskAlias, "%s%s",
QEMU_DRIVE_HOST_PREFIX, disk->info.alias) < 0) ||
(virAsprintf(&nbd_dest, "nbd:%s:%d:exportname=%s",
hoststr, port, diskAlias) < 0))
goto cleanup;
if (qemuDomainObjEnterMonitorAsync(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
qemuBlockJobSyncBegin(disk);
/* Force "raw" format for NBD export */
mon_ret = qemuMonitorDriveMirror(priv->mon, diskAlias, nbd_dest,
"raw", speed, 0, 0, mirror_flags);
VIR_FREE(diskAlias);
VIR_FREE(nbd_dest);
if (qemuDomainObjExitMonitor(driver, vm) < 0 || mon_ret < 0) {
qemuBlockJobSyncEnd(driver, vm, disk);
goto cleanup;
}
diskPriv->migrating = true;
if (virDomainSaveStatus(driver->xmlopt, cfg->stateDir, vm, driver->caps) < 0) {
VIR_WARN("Failed to save status on vm %s", vm->def->name);
goto cleanup;
}
}
while ((rv = qemuMigrationDriveMirrorReady(driver, vm)) != 1) {
if (rv < 0)
goto cleanup;
if (priv->job.abortJob) {
priv->job.current->type = VIR_DOMAIN_JOB_CANCELLED;
virReportError(VIR_ERR_OPERATION_ABORTED, _("%s: %s"),
qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
_("canceled by client"));
goto cleanup;
}
if (dconn && virConnectIsAlive(dconn) <= 0) {
virReportError(VIR_ERR_OPERATION_FAILED, "%s",
_("Lost connection to destination host"));
goto cleanup;
}
if (virDomainObjWait(vm) < 0)
goto cleanup;
}
/* Okay, all disks are ready. Modify migrate_flags */
*migrate_flags &= ~(QEMU_MONITOR_MIGRATE_NON_SHARED_DISK |
QEMU_MONITOR_MIGRATE_NON_SHARED_INC);
ret = 0;
cleanup:
virObjectUnref(cfg);
VIR_FREE(diskAlias);
VIR_FREE(nbd_dest);
VIR_FREE(hoststr);
return ret;
}
/**
* qemuMigrationIsAllowedHostdev:
* @def: domain definition
*
* Checks that @def does not contain any host devices unsupported accross
* migrations. Returns true if the vm is allowed to migrate.
*/
static bool
qemuMigrationIsAllowedHostdev(const virDomainDef *def)
{
size_t i;
/* Migration with USB host devices is allowed, all other devices are
* forbidden. */
for (i = 0; i < def->nhostdevs; i++) {
virDomainHostdevDefPtr hostdev = def->hostdevs[i];
if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS ||
hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("domain has assigned non-USB host devices"));
return false;
}
}
return true;
}
/**
* qemuMigrationIsAllowed:
* @driver: qemu driver struct
* @vm: domain object
* @remote: migration is remote
* @flags: migration flags (see struct virDomainMigrateFlags)
*
* Validates that the configuration of @vm can be migrated in various
* situations. If @remote is true, the migration happens to remote host. @flags
* is used to check various special migration types according to the request.
*
* Returns true if migration is supported. Reports libvirt error and returns
* false otherwise.
*/
bool
qemuMigrationIsAllowed(virQEMUDriverPtr driver,
virDomainObjPtr vm,
bool remote,
unsigned int flags)
{
int nsnapshots;
int pauseReason;
size_t i;
/* perform these checks only when migrating to remote hosts */
if (remote) {
nsnapshots = virDomainSnapshotObjListNum(vm->snapshots, NULL, 0);
if (nsnapshots < 0)
return false;
if (nsnapshots > 0) {
virReportError(VIR_ERR_OPERATION_INVALID,
_("cannot migrate domain with %d snapshots"),
nsnapshots);
return false;
}
/* cancel migration if disk I/O error is emitted while migrating */
if (flags & VIR_MIGRATE_ABORT_ON_ERROR &&
!(flags & VIR_MIGRATE_OFFLINE) &&
virDomainObjGetState(vm, &pauseReason) == VIR_DOMAIN_PAUSED &&
pauseReason == VIR_DOMAIN_PAUSED_IOERROR) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("cannot migrate domain with I/O error"));
return false;
}
}
/* following checks don't make sense for offline migration */
if (!(flags & VIR_MIGRATE_OFFLINE)) {
if (qemuProcessAutoDestroyActive(driver, vm)) {
virReportError(VIR_ERR_OPERATION_INVALID,
"%s", _("domain is marked for auto destroy"));
return false;
}
if (qemuDomainHasBlockjob(vm, false)) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("domain has an active block job"));
return false;
}
if (!qemuMigrationIsAllowedHostdev(vm->def))
return false;
if (vm->def->cpu && vm->def->cpu->mode != VIR_CPU_MODE_HOST_PASSTHROUGH) {
for (i = 0; i < vm->def->cpu->nfeatures; i++) {
virCPUFeatureDefPtr feature = &vm->def->cpu->features[i];
if (feature->policy != VIR_CPU_FEATURE_REQUIRE)
continue;
/* QEMU blocks migration and save with invariant TSC enabled */
if (STREQ(feature->name, "invtsc")) {
virReportError(VIR_ERR_OPERATION_INVALID,
_("domain has CPU feature: %s"),
feature->name);
return false;
}
}
}
/* Verify that memory device config can be transferred reliably */
for (i = 0; i < vm->def->nmems; i++) {
virDomainMemoryDefPtr mem = vm->def->mems[i];
if (mem->model == VIR_DOMAIN_MEMORY_MODEL_DIMM &&
mem->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("domain's dimm info lacks slot ID "
"or base address"));
return false;
}
}
}
return true;
}
static bool
qemuMigrationIsSafe(virDomainDefPtr def,
size_t nmigrate_disks,
const char **migrate_disks)
{
size_t i;
for (i = 0; i < def->ndisks; i++) {
virDomainDiskDefPtr disk = def->disks[i];
const char *src = virDomainDiskGetSource(disk);
/* Our code elsewhere guarantees shared disks are either readonly (in
* which case cache mode doesn't matter) or used with cache=none */
if (qemuMigrateDisk(disk, nmigrate_disks, migrate_disks) &&
disk->cachemode != VIR_DOMAIN_DISK_CACHE_DISABLE) {
int rc;
if (virDomainDiskGetType(disk) == VIR_STORAGE_TYPE_FILE) {
if ((rc = virFileIsSharedFS(src)) < 0)
return false;
else if (rc == 0)
continue;
if ((rc = virStorageFileIsClusterFS(src)) < 0)
return false;
else if (rc == 1)
continue;
} else if (disk->src->type == VIR_STORAGE_TYPE_NETWORK &&
disk->src->protocol == VIR_STORAGE_NET_PROTOCOL_RBD) {
continue;
}
virReportError(VIR_ERR_MIGRATE_UNSAFE, "%s",
_("Migration may lead to data corruption if disks"
" use cache != none"));
return false;
}
}
return true;
}
/** qemuMigrationSetOffline
* Pause domain for non-live migration.
*/
int
qemuMigrationSetOffline(virQEMUDriverPtr driver,
virDomainObjPtr vm)
{
int ret;
VIR_DEBUG("driver=%p vm=%p", driver, vm);
ret = qemuProcessStopCPUs(driver, vm, VIR_DOMAIN_PAUSED_MIGRATION,
QEMU_ASYNC_JOB_MIGRATION_OUT);
if (ret == 0) {
virObjectEventPtr event;
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_SUSPENDED,
VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED);
qemuDomainEventQueue(driver, event);
}
return ret;
}
void
qemuMigrationPostcopyFailed(virQEMUDriverPtr driver,
virDomainObjPtr vm)
{
virDomainState state;
int reason;
state = virDomainObjGetState(vm, &reason);
if (state != VIR_DOMAIN_PAUSED &&
state != VIR_DOMAIN_RUNNING)
return;
if (state == VIR_DOMAIN_PAUSED &&
reason == VIR_DOMAIN_PAUSED_POSTCOPY_FAILED)
return;
VIR_WARN("Migration of domain %s failed during post-copy; "
"leaving the domain paused", vm->def->name);
if (state == VIR_DOMAIN_RUNNING) {
virObjectEventPtr event;
if (qemuProcessStopCPUs(driver, vm,
VIR_DOMAIN_PAUSED_POSTCOPY_FAILED,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0) {
VIR_WARN("Unable to pause guest CPUs for %s", vm->def->name);
return;
}
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_SUSPENDED,
VIR_DOMAIN_EVENT_SUSPENDED_POSTCOPY_FAILED);
qemuDomainEventQueue(driver, event);
} else {
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED,
VIR_DOMAIN_PAUSED_POSTCOPY_FAILED);
}
}
static int
qemuMigrationSetOption(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMonitorMigrationCaps capability,
bool state,
qemuDomainAsyncJob job)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int ret;
if (qemuDomainObjEnterMonitorAsync(driver, vm, job) < 0)
return -1;
ret = qemuMonitorGetMigrationCapability(priv->mon, capability);
if (ret < 0) {
goto cleanup;
} else if (ret == 0 && !state) {
/* Unsupported but we want it off anyway */
goto cleanup;
} else if (ret == 0) {
if (job == QEMU_ASYNC_JOB_MIGRATION_IN) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED,
_("Migration option '%s' is not supported by "
"target QEMU binary"),
qemuMonitorMigrationCapsTypeToString(capability));
} else {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED,
_("Migration option '%s' is not supported by "
"source QEMU binary"),
qemuMonitorMigrationCapsTypeToString(capability));
}
ret = -1;
goto cleanup;
}
ret = qemuMonitorSetMigrationCapability(priv->mon, capability, state);
cleanup:
if (qemuDomainObjExitMonitor(driver, vm) < 0)
ret = -1;
return ret;
}
static int
qemuMigrationSetPostCopy(virQEMUDriverPtr driver,
virDomainObjPtr vm,
bool state,
qemuDomainAsyncJob job)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (qemuMigrationSetOption(driver, vm,
QEMU_MONITOR_MIGRATION_CAPS_POSTCOPY,
state, job) < 0)
return -1;
priv->job.postcopyEnabled = state;
return 0;
}
static int
qemuMigrationWaitForSpice(virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
bool wait_for_spice = false;
size_t i = 0;
if (!virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_SEAMLESS_MIGRATION) ||
!priv->job.spiceMigration)
return 0;
for (i = 0; i < vm->def->ngraphics; i++) {
if (vm->def->graphics[i]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
wait_for_spice = true;
break;
}
}
if (!wait_for_spice)
return 0;
VIR_DEBUG("Waiting for SPICE to finish migration");
while (!priv->job.spiceMigrated && !priv->job.abortJob) {
if (virDomainObjWait(vm) < 0)
return -1;
}
return 0;
}
static void
qemuMigrationUpdateJobType(qemuDomainJobInfoPtr jobInfo)
{
switch ((qemuMonitorMigrationStatus) jobInfo->stats.status) {
case QEMU_MONITOR_MIGRATION_STATUS_COMPLETED:
jobInfo->type = VIR_DOMAIN_JOB_COMPLETED;
break;
case QEMU_MONITOR_MIGRATION_STATUS_INACTIVE:
jobInfo->type = VIR_DOMAIN_JOB_NONE;
break;
case QEMU_MONITOR_MIGRATION_STATUS_ERROR:
jobInfo->type = VIR_DOMAIN_JOB_FAILED;
break;
case QEMU_MONITOR_MIGRATION_STATUS_CANCELLED:
jobInfo->type = VIR_DOMAIN_JOB_CANCELLED;
break;
case QEMU_MONITOR_MIGRATION_STATUS_SETUP:
case QEMU_MONITOR_MIGRATION_STATUS_ACTIVE:
case QEMU_MONITOR_MIGRATION_STATUS_POSTCOPY:
case QEMU_MONITOR_MIGRATION_STATUS_CANCELLING:
case QEMU_MONITOR_MIGRATION_STATUS_LAST:
break;
}
}
int
qemuMigrationFetchJobStatus(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuDomainAsyncJob asyncJob,
qemuDomainJobInfoPtr jobInfo)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int rv;
if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
return -1;
memset(&jobInfo->stats, 0, sizeof(jobInfo->stats));
rv = qemuMonitorGetMigrationStats(priv->mon, &jobInfo->stats);
if (qemuDomainObjExitMonitor(driver, vm) < 0 || rv < 0)
return -1;
qemuMigrationUpdateJobType(jobInfo);
return qemuDomainJobInfoUpdateTime(jobInfo);
}
static const char *
qemuMigrationJobName(virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
switch (priv->job.asyncJob) {
case QEMU_ASYNC_JOB_MIGRATION_OUT:
return _("migration job");
case QEMU_ASYNC_JOB_SAVE:
return _("domain save job");
case QEMU_ASYNC_JOB_DUMP:
return _("domain core dump job");
default:
return _("job");
}
}
static int
qemuMigrationUpdateJobStatus(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuDomainAsyncJob asyncJob)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
qemuDomainJobInfoPtr jobInfo = priv->job.current;
qemuDomainJobInfo newInfo = *jobInfo;
if (qemuMigrationFetchJobStatus(driver, vm, asyncJob, &newInfo) < 0)
return -1;
*jobInfo = newInfo;
return 0;
}
static int
qemuMigrationCheckJobStatus(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuDomainAsyncJob asyncJob,
bool updateJobStats)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
qemuDomainJobInfoPtr jobInfo = priv->job.current;
bool events = virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATION_EVENT);
if (events)
qemuMigrationUpdateJobType(jobInfo);
else if (qemuMigrationUpdateJobStatus(driver, vm, asyncJob) < 0)
return -1;
switch (jobInfo->type) {
case VIR_DOMAIN_JOB_NONE:
virReportError(VIR_ERR_OPERATION_FAILED, _("%s: %s"),
qemuMigrationJobName(vm), _("is not active"));
return -1;
case VIR_DOMAIN_JOB_FAILED:
virReportError(VIR_ERR_OPERATION_FAILED, _("%s: %s"),
qemuMigrationJobName(vm), _("unexpectedly failed"));
return -1;
case VIR_DOMAIN_JOB_CANCELLED:
virReportError(VIR_ERR_OPERATION_ABORTED, _("%s: %s"),
qemuMigrationJobName(vm), _("canceled by client"));
return -1;
case VIR_DOMAIN_JOB_COMPLETED:
/* Fetch statistics of a completed migration */
if (events && updateJobStats &&
qemuMigrationUpdateJobStatus(driver, vm, asyncJob) < 0)
return -1;
break;
case VIR_DOMAIN_JOB_BOUNDED:
case VIR_DOMAIN_JOB_UNBOUNDED:
case VIR_DOMAIN_JOB_LAST:
break;
}
return 0;
}
enum qemuMigrationCompletedFlags {
QEMU_MIGRATION_COMPLETED_ABORT_ON_ERROR = (1 << 0),
QEMU_MIGRATION_COMPLETED_CHECK_STORAGE = (1 << 1),
QEMU_MIGRATION_COMPLETED_UPDATE_STATS = (1 << 2),
QEMU_MIGRATION_COMPLETED_POSTCOPY = (1 << 3),
};
/**
* Returns 1 if migration completed successfully,
* 0 if the domain is still being migrated,
* -1 migration failed,
* -2 something else failed, we need to cancel migration.
*/
static int
qemuMigrationCompleted(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuDomainAsyncJob asyncJob,
virConnectPtr dconn,
unsigned int flags)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
qemuDomainJobInfoPtr jobInfo = priv->job.current;
int pauseReason;
bool updateStats = !!(flags & QEMU_MIGRATION_COMPLETED_UPDATE_STATS);
if (qemuMigrationCheckJobStatus(driver, vm, asyncJob, updateStats) < 0)
goto error;
if (flags & QEMU_MIGRATION_COMPLETED_CHECK_STORAGE &&
qemuMigrationDriveMirrorReady(driver, vm) < 0)
goto error;
if (flags & QEMU_MIGRATION_COMPLETED_ABORT_ON_ERROR &&
virDomainObjGetState(vm, &pauseReason) == VIR_DOMAIN_PAUSED &&
pauseReason == VIR_DOMAIN_PAUSED_IOERROR) {
virReportError(VIR_ERR_OPERATION_FAILED, _("%s: %s"),
qemuMigrationJobName(vm), _("failed due to I/O error"));
goto error;
}
if (dconn && virConnectIsAlive(dconn) <= 0) {
virReportError(VIR_ERR_OPERATION_FAILED, "%s",
_("Lost connection to destination host"));
goto error;
}
/* In case of postcopy the source considers migration completed at the
* moment it switched from active to postcopy-active state. The destination
* will continue waiting until the migrate state changes to completed.
*/
if (flags & QEMU_MIGRATION_COMPLETED_POSTCOPY &&
jobInfo->type == VIR_DOMAIN_JOB_UNBOUNDED &&
jobInfo->stats.status == QEMU_MONITOR_MIGRATION_STATUS_POSTCOPY) {
VIR_DEBUG("Migration switched to post-copy");
if (updateStats &&
qemuMigrationUpdateJobStatus(driver, vm, asyncJob) < 0)
goto error;
return 1;
}
if (jobInfo->type == VIR_DOMAIN_JOB_COMPLETED)
return 1;
else
return 0;
error:
if (jobInfo->type == VIR_DOMAIN_JOB_UNBOUNDED) {
/* The migration was aborted by us rather than QEMU itself. */
jobInfo->type = VIR_DOMAIN_JOB_FAILED;
return -2;
} else if (jobInfo->type == VIR_DOMAIN_JOB_COMPLETED) {
jobInfo->type = VIR_DOMAIN_JOB_FAILED;
return -1;
} else {
return -1;
}
}
/* Returns 0 on success, -2 when migration needs to be cancelled, or -1 when
* QEMU reports failed migration.
*/
static int
qemuMigrationWaitForCompletion(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuDomainAsyncJob asyncJob,
virConnectPtr dconn,
bool abort_on_error,
bool storage)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
qemuDomainJobInfoPtr jobInfo = priv->job.current;
bool events = virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATION_EVENT);
unsigned int flags;
int rv;
flags = QEMU_MIGRATION_COMPLETED_UPDATE_STATS |
QEMU_MIGRATION_COMPLETED_POSTCOPY;
if (abort_on_error)
flags |= QEMU_MIGRATION_COMPLETED_ABORT_ON_ERROR;
if (storage)
flags |= QEMU_MIGRATION_COMPLETED_CHECK_STORAGE;
jobInfo->type = VIR_DOMAIN_JOB_UNBOUNDED;
while ((rv = qemuMigrationCompleted(driver, vm, asyncJob,
dconn, flags)) != 1) {
if (rv < 0)
return rv;
if (events) {
if (virDomainObjWait(vm) < 0) {
jobInfo->type = VIR_DOMAIN_JOB_FAILED;
return -2;
}
} else {
/* Poll every 50ms for progress & to allow cancellation */
struct timespec ts = { .tv_sec = 0, .tv_nsec = 50 * 1000 * 1000ull };
virObjectUnlock(vm);
nanosleep(&ts, NULL);
virObjectLock(vm);
}
}
qemuDomainJobInfoUpdateDowntime(jobInfo);
VIR_FREE(priv->job.completed);
if (VIR_ALLOC(priv->job.completed) == 0)
*priv->job.completed = *jobInfo;
return 0;
}
static int
qemuMigrationWaitForDestCompletion(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuDomainAsyncJob asyncJob,
bool postcopy)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
unsigned int flags = 0;
int rv;
if (!virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATION_EVENT))
return 0;
VIR_DEBUG("Waiting for incoming migration to complete");
if (postcopy)
flags = QEMU_MIGRATION_COMPLETED_POSTCOPY;
while ((rv = qemuMigrationCompleted(driver, vm, asyncJob,
NULL, flags)) != 1) {
if (rv < 0 || virDomainObjWait(vm) < 0)
return -1;
}
return 0;
}
static int
qemuDomainMigrateGraphicsRelocate(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMigrationCookiePtr cookie,
const char *graphicsuri)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int ret = -1;
const char *listenAddress = NULL;
virSocketAddr addr;
virURIPtr uri = NULL;
int type = -1;
int port = -1;
int tlsPort = -1;
const char *tlsSubject = NULL;
if (!cookie || (!cookie->graphics && !graphicsuri))
return 0;
if (graphicsuri && !(uri = virURIParse(graphicsuri)))
goto cleanup;
if (cookie->graphics) {
type = cookie->graphics->type;
listenAddress = cookie->graphics->listen;
if (!listenAddress ||
(virSocketAddrParse(&addr, listenAddress, AF_UNSPEC) > 0 &&
virSocketAddrIsWildcard(&addr)))
listenAddress = cookie->remoteHostname;
port = cookie->graphics->port;
tlsPort = cookie->graphics->tlsPort;
tlsSubject = cookie->graphics->tlsSubject;
}
if (uri) {
size_t i;
if ((type = virDomainGraphicsTypeFromString(uri->scheme)) < 0) {
virReportError(VIR_ERR_INVALID_ARG,
_("unknown graphics type %s"), uri->scheme);
goto cleanup;
}
if (uri->server)
listenAddress = uri->server;
if (uri->port > 0)
port = uri->port;
for (i = 0; i < uri->paramsCount; i++) {
virURIParamPtr param = uri->params + i;
if (STRCASEEQ(param->name, "tlsPort")) {
if (virStrToLong_i(param->value, NULL, 10, &tlsPort) < 0) {
virReportError(VIR_ERR_INVALID_ARG,
_("invalid tlsPort number: %s"),
param->value);
goto cleanup;
}
} else if (STRCASEEQ(param->name, "tlsSubject")) {
tlsSubject = param->value;
}
}
}
/* QEMU doesn't support VNC relocation yet, so
* skip it to avoid generating an error
*/
if (type != VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
ret = 0;
goto cleanup;
}
if (qemuDomainObjEnterMonitorAsync(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_OUT) == 0) {
ret = qemuMonitorGraphicsRelocate(priv->mon, type, listenAddress,
port, tlsPort, tlsSubject);
priv->job.spiceMigration = !ret;
if (qemuDomainObjExitMonitor(driver, vm) < 0)
ret = -1;
}
cleanup:
virURIFree(uri);
return ret;
}
static int
qemuDomainMigrateOPDRelocate(virQEMUDriverPtr driver ATTRIBUTE_UNUSED,
virDomainObjPtr vm,
qemuMigrationCookiePtr cookie)
{
virDomainNetDefPtr netptr;
int ret = -1;
size_t i;
for (i = 0; i < cookie->network->nnets; i++) {
netptr = vm->def->nets[i];
switch (cookie->network->net[i].vporttype) {
case VIR_NETDEV_VPORT_PROFILE_NONE:
case VIR_NETDEV_VPORT_PROFILE_8021QBG:
case VIR_NETDEV_VPORT_PROFILE_8021QBH:
break;
case VIR_NETDEV_VPORT_PROFILE_OPENVSWITCH:
if (virNetDevOpenvswitchSetMigrateData(cookie->network->net[i].portdata,
netptr->ifname) != 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Unable to run command to set OVS port data for "
"interface %s"), netptr->ifname);
goto cleanup;
}
break;
default:
break;
}
}
ret = 0;
cleanup:
return ret;
}
int
qemuMigrationCheckIncoming(virQEMUCapsPtr qemuCaps,
const char *migrateFrom)
{
if (STRPREFIX(migrateFrom, "rdma")) {
if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_MIGRATE_RDMA)) {
virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s",
_("incoming RDMA migration is not supported "
"with this QEMU binary"));
return -1;
}
} else if (!STRPREFIX(migrateFrom, "tcp") &&
!STRPREFIX(migrateFrom, "exec") &&
!STRPREFIX(migrateFrom, "fd") &&
!STRPREFIX(migrateFrom, "unix") &&
STRNEQ(migrateFrom, "stdio")) {
virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s",
_("unknown migration protocol"));
return -1;
}
return 0;
}
char *
qemuMigrationIncomingURI(const char *migrateFrom,
int migrateFd)
{
char *uri = NULL;
if (STREQ(migrateFrom, "stdio"))
ignore_value(virAsprintf(&uri, "fd:%d", migrateFd));
else
ignore_value(VIR_STRDUP(uri, migrateFrom));
return uri;
}
int
qemuMigrationRunIncoming(virQEMUDriverPtr driver,
virDomainObjPtr vm,
const char *uri,
qemuDomainAsyncJob asyncJob)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int ret = -1;
int rv;
VIR_DEBUG("Setting up incoming migration with URI %s", uri);
if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
return -1;
rv = qemuMonitorMigrateIncoming(priv->mon, uri);
if (qemuDomainObjExitMonitor(driver, vm) < 0 || rv < 0)
goto cleanup;
if (asyncJob == QEMU_ASYNC_JOB_MIGRATION_IN) {
/* qemuMigrationWaitForDestCompletion is called from the Finish phase */
ret = 0;
goto cleanup;
}
if (qemuMigrationWaitForDestCompletion(driver, vm, asyncJob, false) < 0)
goto cleanup;
ret = 0;
cleanup:
return ret;
}
/* This is called for outgoing non-p2p migrations when a connection to the
* client which initiated the migration was closed but we were waiting for it
* to follow up with the next phase, that is, in between
* qemuDomainMigrateBegin3 and qemuDomainMigratePerform3 or
* qemuDomainMigratePerform3 and qemuDomainMigrateConfirm3.
*/
static virDomainObjPtr
qemuMigrationCleanup(virDomainObjPtr vm,
virConnectPtr conn,
void *opaque)
{
virQEMUDriverPtr driver = opaque;
qemuDomainObjPrivatePtr priv = vm->privateData;
VIR_DEBUG("vm=%s, conn=%p, asyncJob=%s, phase=%s",
vm->def->name, conn,
qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
qemuDomainAsyncJobPhaseToString(priv->job.asyncJob,
priv->job.phase));
if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_OUT))
goto cleanup;
VIR_DEBUG("The connection which started outgoing migration of domain %s"
" was closed; canceling the migration",
vm->def->name);
switch ((qemuMigrationJobPhase) priv->job.phase) {
case QEMU_MIGRATION_PHASE_BEGIN3:
/* just forget we were about to migrate */
qemuDomainObjDiscardAsyncJob(driver, vm);
break;
case QEMU_MIGRATION_PHASE_PERFORM3_DONE:
VIR_WARN("Migration of domain %s finished but we don't know if the"
" domain was successfully started on destination or not",
vm->def->name);
/* clear the job and let higher levels decide what to do */
qemuDomainObjDiscardAsyncJob(driver, vm);
break;
case QEMU_MIGRATION_PHASE_PERFORM3:
/* cannot be seen without an active migration API; unreachable */
case QEMU_MIGRATION_PHASE_CONFIRM3:
case QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED:
/* all done; unreachable */
case QEMU_MIGRATION_PHASE_PREPARE:
case QEMU_MIGRATION_PHASE_FINISH2:
case QEMU_MIGRATION_PHASE_FINISH3:
/* incoming migration; unreachable */
case QEMU_MIGRATION_PHASE_PERFORM2:
/* single phase outgoing migration; unreachable */
case QEMU_MIGRATION_PHASE_NONE:
case QEMU_MIGRATION_PHASE_LAST:
/* unreachable */
;
}
cleanup:
return vm;
}
/* The caller is supposed to lock the vm and start a migration job. */
static char *
qemuMigrationBeginPhase(virQEMUDriverPtr driver,
virDomainObjPtr vm,
const char *xmlin,
const char *dname,
char **cookieout,
int *cookieoutlen,
size_t nmigrate_disks,
const char **migrate_disks,
unsigned long flags)
{
char *rv = NULL;
qemuMigrationCookiePtr mig = NULL;
virDomainDefPtr def = NULL;
qemuDomainObjPrivatePtr priv = vm->privateData;
virCapsPtr caps = NULL;
unsigned int cookieFlags = QEMU_MIGRATION_COOKIE_LOCKSTATE;
VIR_DEBUG("driver=%p, vm=%p, xmlin=%s, dname=%s,"
" cookieout=%p, cookieoutlen=%p,"
" nmigrate_disks=%zu, migrate_disks=%p, flags=%lx",
driver, vm, NULLSTR(xmlin), NULLSTR(dname),
cookieout, cookieoutlen, nmigrate_disks,
migrate_disks, flags);
if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
goto cleanup;
/* Only set the phase if we are inside QEMU_ASYNC_JOB_MIGRATION_OUT.
* Otherwise we will start the async job later in the perform phase losing
* change protection.
*/
if (priv->job.asyncJob == QEMU_ASYNC_JOB_MIGRATION_OUT)
qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_BEGIN3);
if (!qemuMigrationIsAllowed(driver, vm, true, flags))
goto cleanup;
if (!(flags & VIR_MIGRATE_UNSAFE) &&
!qemuMigrationIsSafe(vm->def, nmigrate_disks, migrate_disks))
goto cleanup;
if (flags & VIR_MIGRATE_POSTCOPY &&
(!(flags & VIR_MIGRATE_LIVE) ||
flags & VIR_MIGRATE_PAUSED)) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("post-copy migration is not supported with non-live "
"or paused migration"));
goto cleanup;
}
if (flags & (VIR_MIGRATE_NON_SHARED_DISK | VIR_MIGRATE_NON_SHARED_INC)) {
bool has_drive_mirror = virQEMUCapsGet(priv->qemuCaps,
QEMU_CAPS_DRIVE_MIRROR);
if (nmigrate_disks) {
if (has_drive_mirror) {
size_t i, j;
/* Check user requested only known disk targets. */
for (i = 0; i < nmigrate_disks; i++) {
for (j = 0; j < vm->def->ndisks; j++) {
if (STREQ(vm->def->disks[j]->dst, migrate_disks[i]))
break;
}
if (j == vm->def->ndisks) {
virReportError(VIR_ERR_INVALID_ARG,
_("disk target %s not found"),
migrate_disks[i]);
goto cleanup;
}
}
if (flags & VIR_MIGRATE_TUNNELLED) {
virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s",
_("Selecting disks to migrate is not "
"implemented for tunnelled migration"));
goto cleanup;
}
} else {
virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s",
_("qemu does not support drive-mirror command"));
goto cleanup;
}
}
if (has_drive_mirror) {
/* TODO support NBD for TUNNELLED migration */
if (flags & VIR_MIGRATE_TUNNELLED) {
VIR_WARN("NBD in tunnelled migration is currently not supported");
} else {
cookieFlags |= QEMU_MIGRATION_COOKIE_NBD;
priv->nbdPort = 0;
}
}
}
if (virDomainDefHasMemoryHotplug(vm->def) ||
((flags & VIR_MIGRATE_PERSIST_DEST) &&
vm->newDef && virDomainDefHasMemoryHotplug(vm->newDef)))
cookieFlags |= QEMU_MIGRATION_COOKIE_MEMORY_HOTPLUG;
if (!(mig = qemuMigrationEatCookie(driver, vm, NULL, 0, 0)))
goto cleanup;
if (qemuMigrationBakeCookie(mig, driver, vm,
cookieout, cookieoutlen,
cookieFlags) < 0)
goto cleanup;
if (flags & VIR_MIGRATE_OFFLINE) {
if (flags & (VIR_MIGRATE_NON_SHARED_DISK |
VIR_MIGRATE_NON_SHARED_INC)) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("offline migration cannot handle "
"non-shared storage"));
goto cleanup;
}
if (!(flags & VIR_MIGRATE_PERSIST_DEST)) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("offline migration must be specified with "
"the persistent flag set"));
goto cleanup;
}
if (flags & VIR_MIGRATE_TUNNELLED) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("tunnelled offline migration does not "
"make sense"));
goto cleanup;
}
}
if (xmlin) {
if (!(def = virDomainDefParseString(xmlin, caps, driver->xmlopt,
VIR_DOMAIN_DEF_PARSE_INACTIVE)))
goto cleanup;
if (!qemuDomainDefCheckABIStability(driver, vm->def, def))
goto cleanup;
rv = qemuDomainDefFormatLive(driver, def, false, true);
} else {
rv = qemuDomainDefFormatLive(driver, vm->def, false, true);
}
cleanup:
qemuMigrationCookieFree(mig);
virObjectUnref(caps);
virDomainDefFree(def);
return rv;
}
char *
qemuMigrationBegin(virConnectPtr conn,
virDomainObjPtr vm,
const char *xmlin,
const char *dname,
char **cookieout,
int *cookieoutlen,
size_t nmigrate_disks,
const char **migrate_disks,
unsigned long flags)
{
virQEMUDriverPtr driver = conn->privateData;
char *xml = NULL;
qemuDomainAsyncJob asyncJob;
if ((flags & VIR_MIGRATE_CHANGE_PROTECTION)) {
if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
asyncJob = QEMU_ASYNC_JOB_MIGRATION_OUT;
} else {
if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_MODIFY) < 0)
goto cleanup;
asyncJob = QEMU_ASYNC_JOB_NONE;
}
qemuMigrationStoreDomainState(vm);
if (!virDomainObjIsActive(vm) && !(flags & VIR_MIGRATE_OFFLINE)) {
virReportError(VIR_ERR_OPERATION_INVALID,
"%s", _("domain is not running"));
goto endjob;
}
/* Check if there is any ejected media.
* We don't want to require them on the destination.
*/
if (!(flags & VIR_MIGRATE_OFFLINE) &&
qemuDomainCheckEjectableMedia(driver, vm, asyncJob) < 0)
goto endjob;
if (!(xml = qemuMigrationBeginPhase(driver, vm, xmlin, dname,
cookieout, cookieoutlen,
nmigrate_disks, migrate_disks, flags)))
goto endjob;
if ((flags & VIR_MIGRATE_CHANGE_PROTECTION)) {
/* We keep the job active across API calls until the confirm() call.
* This prevents any other APIs being invoked while migration is taking
* place.
*/
if (virCloseCallbacksSet(driver->closeCallbacks, vm, conn,
qemuMigrationCleanup) < 0) {
VIR_FREE(xml);
goto endjob;
}
qemuMigrationJobContinue(vm);
} else {
goto endjob;
}
cleanup:
virDomainObjEndAPI(&vm);
return xml;
endjob:
if (flags & VIR_MIGRATE_CHANGE_PROTECTION)
qemuMigrationJobFinish(driver, vm);
else
qemuDomainObjEndJob(driver, vm);
goto cleanup;
}
/* Prepare is the first step, and it runs on the destination host.
*/
static void
qemuMigrationPrepareCleanup(virQEMUDriverPtr driver,
virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
VIR_DEBUG("driver=%p, vm=%s, job=%s, asyncJob=%s",
driver,
vm->def->name,
qemuDomainJobTypeToString(priv->job.active),
qemuDomainAsyncJobTypeToString(priv->job.asyncJob));
virPortAllocatorRelease(driver->migrationPorts, priv->migrationPort);
priv->migrationPort = 0;
if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_IN))
return;
qemuDomainObjDiscardAsyncJob(driver, vm);
}
static qemuProcessIncomingDefPtr
qemuMigrationPrepareIncoming(virDomainObjPtr vm,
bool tunnel,
const char *protocol,
const char *listenAddress,
unsigned short port,
int fd)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
qemuProcessIncomingDefPtr inc = NULL;
char *migrateFrom = NULL;
if (tunnel) {
if (VIR_STRDUP(migrateFrom, "stdio") < 0)
goto cleanup;
} else {
bool encloseAddress = false;
bool hostIPv6Capable = false;
bool qemuIPv6Capable = false;
struct addrinfo *info = NULL;
struct addrinfo hints = { .ai_flags = AI_ADDRCONFIG,
.ai_socktype = SOCK_STREAM };
const char *incFormat;
if (getaddrinfo("::", NULL, &hints, &info) == 0) {
freeaddrinfo(info);
hostIPv6Capable = true;
}
qemuIPv6Capable = virQEMUCapsGet(priv->qemuCaps,
QEMU_CAPS_IPV6_MIGRATION);
if (listenAddress) {
if (virSocketAddrNumericFamily(listenAddress) == AF_INET6) {
if (!qemuIPv6Capable) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("qemu isn't capable of IPv6"));
goto cleanup;
}
if (!hostIPv6Capable) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("host isn't capable of IPv6"));
goto cleanup;
}
/* IPv6 address must be escaped in brackets on the cmd line */
encloseAddress = true;
} else {
/* listenAddress is a hostname or IPv4 */
}
} else if (qemuIPv6Capable && hostIPv6Capable) {
/* Listen on :: instead of 0.0.0.0 if QEMU understands it
* and there is at least one IPv6 address configured
*/
listenAddress = "::";
encloseAddress = true;
} else {
listenAddress = "0.0.0.0";
}
/* QEMU will be started with
* -incoming protocol:[]:port,
* -incoming protocol::port, or
* -incoming protocol::port
*/
if (encloseAddress)
incFormat = "%s:[%s]:%d";
else
incFormat = "%s:%s:%d";
if (virAsprintf(&migrateFrom, incFormat,
protocol, listenAddress, port) < 0)
goto cleanup;
}
inc = qemuProcessIncomingDefNew(priv->qemuCaps, listenAddress,
migrateFrom, fd, NULL);
cleanup:
VIR_FREE(migrateFrom);
return inc;
}
static int
qemuMigrationPrepareAny(virQEMUDriverPtr driver,
virConnectPtr dconn,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
virDomainDefPtr *def,
const char *origname,
virStreamPtr st,
const char *protocol,
unsigned short port,
bool autoPort,
const char *listenAddress,
size_t nmigrate_disks,
const char **migrate_disks,
int nbdPort,
unsigned long flags)
{
virDomainObjPtr vm = NULL;
virObjectEventPtr event = NULL;
int ret = -1;
int dataFD[2] = { -1, -1 };
qemuDomainObjPrivatePtr priv = NULL;
qemuMigrationCookiePtr mig = NULL;
bool tunnel = !!st;
char *xmlout = NULL;
unsigned int cookieFlags;
virCapsPtr caps = NULL;
qemuProcessIncomingDefPtr incoming = NULL;
bool taint_hook = false;
bool stopProcess = false;
bool relabel = false;
int rv;
virNWFilterReadLockFilterUpdates();
if (flags & VIR_MIGRATE_OFFLINE) {
if (flags & (VIR_MIGRATE_NON_SHARED_DISK |
VIR_MIGRATE_NON_SHARED_INC)) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("offline migration cannot handle "
"non-shared storage"));
goto cleanup;
}
if (!(flags & VIR_MIGRATE_PERSIST_DEST)) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("offline migration must be specified with "
"the persistent flag set"));
goto cleanup;
}
if (tunnel) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("tunnelled offline migration does not "
"make sense"));
goto cleanup;
}
cookieFlags = 0;
} else {
cookieFlags = QEMU_MIGRATION_COOKIE_GRAPHICS;
}
if (flags & VIR_MIGRATE_POSTCOPY &&
(!(flags & VIR_MIGRATE_LIVE) ||
flags & VIR_MIGRATE_PAUSED)) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("post-copy migration is not supported with non-live "
"or paused migration"));
goto cleanup;
}
if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
goto cleanup;
if (!qemuMigrationIsAllowedHostdev(*def))
goto cleanup;
/* Let migration hook filter domain XML */
if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
char *xml;
int hookret;
if (!(xml = qemuDomainDefFormatXML(driver, *def,
VIR_DOMAIN_XML_SECURE |
VIR_DOMAIN_XML_MIGRATABLE)))
goto cleanup;
hookret = virHookCall(VIR_HOOK_DRIVER_QEMU, (*def)->name,
VIR_HOOK_QEMU_OP_MIGRATE, VIR_HOOK_SUBOP_BEGIN,
NULL, xml, &xmlout);
VIR_FREE(xml);
if (hookret < 0) {
goto cleanup;
} else if (hookret == 0) {
if (virStringIsEmpty(xmlout)) {
VIR_DEBUG("Migrate hook filter returned nothing; using the"
" original XML");
} else {
virDomainDefPtr newdef;
VIR_DEBUG("Using hook-filtered domain XML: %s", xmlout);
newdef = virDomainDefParseString(xmlout, caps, driver->xmlopt,
VIR_DOMAIN_DEF_PARSE_INACTIVE);
if (!newdef)
goto cleanup;
if (!qemuDomainDefCheckABIStability(driver, *def, newdef)) {
virDomainDefFree(newdef);
goto cleanup;
}
virDomainDefFree(*def);
*def = newdef;
/* We should taint the domain here. However, @vm and therefore
* privateData too are still NULL, so just notice the fact and
* taint it later. */
taint_hook = true;
}
}
}
if (!(vm = virDomainObjListAdd(driver->domains, *def,
driver->xmlopt,
VIR_DOMAIN_OBJ_LIST_ADD_LIVE |
VIR_DOMAIN_OBJ_LIST_ADD_CHECK_LIVE,
NULL)))
goto cleanup;
virObjectRef(vm);
*def = NULL;
priv = vm->privateData;
if (VIR_STRDUP(priv->origname, origname) < 0)
goto cleanup;
if (taint_hook) {
/* Domain XML has been altered by a hook script. */
priv->hookRun = true;
}
if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen,
QEMU_MIGRATION_COOKIE_LOCKSTATE |
QEMU_MIGRATION_COOKIE_NBD |
QEMU_MIGRATION_COOKIE_MEMORY_HOTPLUG)))
goto cleanup;
if (STREQ_NULLABLE(protocol, "rdma") &&
!virMemoryLimitIsSet(vm->def->mem.hard_limit)) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("cannot start RDMA migration with no memory hard "
"limit set"));
goto cleanup;
}
if (qemuMigrationPrecreateStorage(dconn, driver, vm, mig->nbd,
nmigrate_disks, migrate_disks,
!!(flags & VIR_MIGRATE_NON_SHARED_INC)) < 0)
goto cleanup;
if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
goto cleanup;
qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PREPARE);
/* Domain starts inactive, even if the domain XML had an id field. */
vm->def->id = -1;
if (flags & VIR_MIGRATE_OFFLINE)
goto done;
if (tunnel &&
(pipe(dataFD) < 0 || virSetCloseExec(dataFD[1]) < 0)) {
virReportSystemError(errno, "%s",
_("cannot create pipe for tunnelled migration"));
goto stopjob;
}
if (qemuProcessInit(driver, vm, QEMU_ASYNC_JOB_MIGRATION_IN,
true, false) < 0)
goto stopjob;
stopProcess = true;
if (!(incoming = qemuMigrationPrepareIncoming(vm, tunnel, protocol,
listenAddress, port,
dataFD[0])))
goto stopjob;
dataFD[0] = -1; /* the FD is now owned by incoming */
rv = qemuProcessLaunch(dconn, driver, vm, QEMU_ASYNC_JOB_MIGRATION_IN,
incoming, NULL,
VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_START,
VIR_QEMU_PROCESS_START_AUTODESTROY);
if (rv < 0) {
if (rv == -2)
relabel = true;
goto stopjob;
}
relabel = true;
if (tunnel) {
if (virFDStreamOpen(st, dataFD[1]) < 0) {
virReportSystemError(errno, "%s",
_("cannot pass pipe for tunnelled migration"));
goto stopjob;
}
dataFD[1] = -1; /* 'st' owns the FD now & will close it */
}
if (qemuMigrationSetOption(driver, vm,
QEMU_MONITOR_MIGRATION_CAPS_XBZRLE,
flags & VIR_MIGRATE_COMPRESSED,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
goto stopjob;
if (STREQ_NULLABLE(protocol, "rdma") &&
virProcessSetMaxMemLock(vm->pid, vm->def->mem.hard_limit << 10) < 0) {
goto stopjob;
}
if (qemuMigrationSetOption(driver, vm,
QEMU_MONITOR_MIGRATION_CAPS_RDMA_PIN_ALL,
flags & VIR_MIGRATE_RDMA_PIN_ALL,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
goto stopjob;
if (qemuMigrationSetPostCopy(driver, vm,
flags & VIR_MIGRATE_POSTCOPY,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
goto stopjob;
if (mig->nbd &&
flags & (VIR_MIGRATE_NON_SHARED_DISK | VIR_MIGRATE_NON_SHARED_INC) &&
virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_NBD_SERVER)) {
if (qemuMigrationStartNBDServer(driver, vm, incoming->address,
nmigrate_disks, migrate_disks,
nbdPort) < 0) {
goto stopjob;
}
cookieFlags |= QEMU_MIGRATION_COOKIE_NBD;
}
if (mig->lockState) {
VIR_DEBUG("Received lockstate %s", mig->lockState);
VIR_FREE(priv->lockState);
priv->lockState = mig->lockState;
mig->lockState = NULL;
} else {
VIR_DEBUG("Received no lockstate");
}
if (incoming->deferredURI &&
qemuMigrationRunIncoming(driver, vm, incoming->deferredURI,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
goto stopjob;
if (qemuProcessFinishStartup(dconn, driver, vm, QEMU_ASYNC_JOB_MIGRATION_IN,
false, VIR_DOMAIN_PAUSED_MIGRATION) < 0)
goto stopjob;
done:
if (qemuMigrationBakeCookie(mig, driver, vm, cookieout,
cookieoutlen, cookieFlags) < 0) {
/* We could tear down the whole guest here, but
* cookie data is (so far) non-critical, so that
* seems a little harsh. We'll just warn for now.
*/
VIR_WARN("Unable to encode migration cookie");
}
if (qemuDomainCleanupAdd(vm, qemuMigrationPrepareCleanup) < 0)
goto stopjob;
if (!(flags & VIR_MIGRATE_OFFLINE)) {
virDomainAuditStart(vm, "migrated", true);
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_STARTED,
VIR_DOMAIN_EVENT_STARTED_MIGRATED);
}
/* We keep the job active across API calls until the finish() call.
* This prevents any other APIs being invoked while incoming
* migration is taking place.
*/
qemuMigrationJobContinue(vm);
if (autoPort)
priv->migrationPort = port;
/* in this case port is not auto selected and we don't need to manage it
* anymore after cookie is baked
*/
if (nbdPort != 0)
priv->nbdPort = 0;
ret = 0;
cleanup:
qemuProcessIncomingDefFree(incoming);
VIR_FREE(xmlout);
VIR_FORCE_CLOSE(dataFD[0]);
VIR_FORCE_CLOSE(dataFD[1]);
if (ret < 0 && priv) {
/* priv is set right after vm is added to the list of domains
* and there is no 'goto cleanup;' in the middle of those */
VIR_FREE(priv->origname);
/* release if port is auto selected which is not the case if
* it is given in parameters
*/
if (nbdPort == 0)
virPortAllocatorRelease(driver->migrationPorts, priv->nbdPort);
priv->nbdPort = 0;
qemuDomainRemoveInactive(driver, vm);
}
virDomainObjEndAPI(&vm);
qemuDomainEventQueue(driver, event);
qemuMigrationCookieFree(mig);
virObjectUnref(caps);
virNWFilterUnlockFilterUpdates();
return ret;
stopjob:
if (stopProcess) {
unsigned int stopFlags = VIR_QEMU_PROCESS_STOP_MIGRATED;
if (!relabel)
stopFlags |= VIR_QEMU_PROCESS_STOP_NO_RELABEL;
virDomainAuditStart(vm, "migrated", false);
qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED,
QEMU_ASYNC_JOB_MIGRATION_IN, stopFlags);
}
qemuMigrationJobFinish(driver, vm);
goto cleanup;
}
/*
* This version starts an empty VM listening on a localhost TCP port, and
* sets up the corresponding virStream to handle the incoming data.
*/
int
qemuMigrationPrepareTunnel(virQEMUDriverPtr driver,
virConnectPtr dconn,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
virStreamPtr st,
virDomainDefPtr *def,
const char *origname,
unsigned long flags)
{
int ret;
VIR_DEBUG("driver=%p, dconn=%p, cookiein=%s, cookieinlen=%d, "
"cookieout=%p, cookieoutlen=%p, st=%p, def=%p, "
"origname=%s, flags=%lx",
driver, dconn, NULLSTR(cookiein), cookieinlen,
cookieout, cookieoutlen, st, *def, origname, flags);
if (st == NULL) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("tunnelled migration requested but NULL stream passed"));
return -1;
}
ret = qemuMigrationPrepareAny(driver, dconn, cookiein, cookieinlen,
cookieout, cookieoutlen, def, origname,
st, NULL, 0, false, NULL, 0, NULL, 0, flags);
return ret;
}
static virURIPtr
qemuMigrationParseURI(const char *uri, bool *wellFormed)
{
char *tmp = NULL;
virURIPtr parsed;
/* For compatibility reasons tcp://... URIs are sent as tcp:...
* We need to transform them to a well-formed URI before parsing. */
if (STRPREFIX(uri, "tcp:") && !STRPREFIX(uri + 4, "//")) {
if (virAsprintf(&tmp, "tcp://%s", uri + 4) < 0)
return NULL;
uri = tmp;
}
parsed = virURIParse(uri);
if (parsed && wellFormed)
*wellFormed = !tmp;
VIR_FREE(tmp);
return parsed;
}
int
qemuMigrationPrepareDirect(virQEMUDriverPtr driver,
virConnectPtr dconn,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
const char *uri_in,
char **uri_out,
virDomainDefPtr *def,
const char *origname,
const char *listenAddress,
size_t nmigrate_disks,
const char **migrate_disks,
int nbdPort,
unsigned long flags)
{
unsigned short port = 0;
bool autoPort = true;
char *hostname = NULL;
int ret = -1;
virURIPtr uri = NULL;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
const char *migrateHost = cfg->migrateHost;
VIR_DEBUG("driver=%p, dconn=%p, cookiein=%s, cookieinlen=%d, "
"cookieout=%p, cookieoutlen=%p, uri_in=%s, uri_out=%p, "
"def=%p, origname=%s, listenAddress=%s, "
"nmigrate_disks=%zu, migrate_disks=%p, nbdPort=%d, flags=%lx",
driver, dconn, NULLSTR(cookiein), cookieinlen,
cookieout, cookieoutlen, NULLSTR(uri_in), uri_out,
*def, origname, NULLSTR(listenAddress),
nmigrate_disks, migrate_disks, nbdPort, flags);
*uri_out = NULL;
/* The URI passed in may be NULL or a string "tcp://somehostname:port".
*
* If the URI passed in is NULL then we allocate a port number
* from our pool of port numbers, and if the migrateHost is configured,
* we return a URI of "tcp://migrateHost:port", otherwise return a URI
* of "tcp://ourhostname:port".
*
* If the URI passed in is not NULL then we try to parse out the
* port number and use that (note that the hostname is assumed
* to be a correct hostname which refers to the target machine).
*/
if (uri_in == NULL) {
bool encloseAddress = false;
const char *incFormat;
if (virPortAllocatorAcquire(driver->migrationPorts, &port) < 0)
goto cleanup;
if (migrateHost != NULL) {
if (virSocketAddrNumericFamily(migrateHost) == AF_INET6)
encloseAddress = true;
if (VIR_STRDUP(hostname, migrateHost) < 0)
goto cleanup;
} else {
if ((hostname = virGetHostname()) == NULL)
goto cleanup;
}
if (STRPREFIX(hostname, "localhost")) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("hostname on destination resolved to localhost,"
" but migration requires an FQDN"));
goto cleanup;
}
/* XXX this really should have been a properly well-formed
* URI, but we can't add in tcp:// now without breaking
* compatibility with old targets. We at least make the
* new targets accept both syntaxes though.
*/
if (encloseAddress)
incFormat = "%s:[%s]:%d";
else
incFormat = "%s:%s:%d";
if (virAsprintf(uri_out, incFormat, "tcp", hostname, port) < 0)
goto cleanup;
} else {
bool well_formed_uri;
if (!(uri = qemuMigrationParseURI(uri_in, &well_formed_uri)))
goto cleanup;
if (uri->scheme == NULL) {
virReportError(VIR_ERR_INVALID_ARG,
_("missing scheme in migration URI: %s"),
uri_in);
goto cleanup;
}
if (STRNEQ(uri->scheme, "tcp") &&
STRNEQ(uri->scheme, "rdma")) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED,
_("unsupported scheme %s in migration URI %s"),
uri->scheme, uri_in);
goto cleanup;
}
if (uri->server == NULL) {
virReportError(VIR_ERR_INVALID_ARG, _("missing host in migration"
" URI: %s"), uri_in);
goto cleanup;
}
if (uri->port == 0) {
if (virPortAllocatorAcquire(driver->migrationPorts, &port) < 0)
goto cleanup;
/* Send well-formed URI only if uri_in was well-formed */
if (well_formed_uri) {
uri->port = port;
if (!(*uri_out = virURIFormat(uri)))
goto cleanup;
} else {
if (virAsprintf(uri_out, "%s:%d", uri_in, port) < 0)
goto cleanup;
}
} else {
port = uri->port;
autoPort = false;
}
}
if (*uri_out)
VIR_DEBUG("Generated uri_out=%s", *uri_out);
ret = qemuMigrationPrepareAny(driver, dconn, cookiein, cookieinlen,
cookieout, cookieoutlen, def, origname,
NULL, uri ? uri->scheme : "tcp",
port, autoPort, listenAddress,
nmigrate_disks, migrate_disks, nbdPort, flags);
cleanup:
virURIFree(uri);
VIR_FREE(hostname);
virObjectUnref(cfg);
if (ret != 0) {
VIR_FREE(*uri_out);
if (autoPort)
virPortAllocatorRelease(driver->migrationPorts, port);
}
return ret;
}
virDomainDefPtr
qemuMigrationPrepareDef(virQEMUDriverPtr driver,
const char *dom_xml,
const char *dname,
char **origname)
{
virCapsPtr caps = NULL;
virDomainDefPtr def;
char *name = NULL;
if (!dom_xml) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("no domain XML passed"));
return NULL;
}
if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
return NULL;
if (!(def = virDomainDefParseString(dom_xml, caps, driver->xmlopt,
VIR_DOMAIN_DEF_PARSE_INACTIVE)))
goto cleanup;
if (dname) {
name = def->name;
if (VIR_STRDUP(def->name, dname) < 0) {
virDomainDefFree(def);
def = NULL;
}
}
cleanup:
virObjectUnref(caps);
if (def && origname)
*origname = name;
else
VIR_FREE(name);
return def;
}
static int
qemuMigrationConfirmPhase(virQEMUDriverPtr driver,
virConnectPtr conn,
virDomainObjPtr vm,
const char *cookiein,
int cookieinlen,
unsigned int flags,
int retcode)
{
qemuMigrationCookiePtr mig;
virObjectEventPtr event;
int rv = -1;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
qemuDomainObjPrivatePtr priv = vm->privateData;
qemuDomainJobInfoPtr jobInfo = NULL;
VIR_DEBUG("driver=%p, conn=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
"flags=%x, retcode=%d",
driver, conn, vm, NULLSTR(cookiein), cookieinlen,
flags, retcode);
virCheckFlags(QEMU_MIGRATION_FLAGS, -1);
qemuMigrationJobSetPhase(driver, vm,
retcode == 0
? QEMU_MIGRATION_PHASE_CONFIRM3
: QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED);
if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen,
QEMU_MIGRATION_COOKIE_STATS)))
goto cleanup;
if (retcode == 0)
jobInfo = priv->job.completed;
else
VIR_FREE(priv->job.completed);
/* Update times with the values sent by the destination daemon */
if (mig->jobInfo && jobInfo) {
int reason;
/* We need to refresh migration statistics after a completed post-copy
* migration since priv->job.completed contains obsolete data from the
* time we switched to post-copy mode.
*/
if (virDomainObjGetState(vm, &reason) == VIR_DOMAIN_PAUSED &&
reason == VIR_DOMAIN_PAUSED_POSTCOPY &&
qemuMigrationFetchJobStatus(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_OUT,
jobInfo) < 0)
VIR_WARN("Could not refresh migration statistics");
qemuDomainJobInfoUpdateTime(jobInfo);
jobInfo->timeDeltaSet = mig->jobInfo->timeDeltaSet;
jobInfo->timeDelta = mig->jobInfo->timeDelta;
jobInfo->stats.downtime_set = mig->jobInfo->stats.downtime_set;
jobInfo->stats.downtime = mig->jobInfo->stats.downtime;
}
if (flags & VIR_MIGRATE_OFFLINE)
goto done;
/* Did the migration go as planned? If yes, kill off the domain object.
* If something failed, resume CPUs, but only if we didn't use post-copy.
*/
if (retcode == 0) {
/* If guest uses SPICE and supports seamless migration we have to hold
* up domain shutdown until SPICE server transfers its data */
qemuMigrationWaitForSpice(vm);
qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_MIGRATED,
QEMU_ASYNC_JOB_MIGRATION_OUT,
VIR_QEMU_PROCESS_STOP_MIGRATED);
virDomainAuditStop(vm, "migrated");
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_STOPPED,
VIR_DOMAIN_EVENT_STOPPED_MIGRATED);
qemuDomainEventQueue(driver, event);
qemuDomainEventEmitJobCompleted(driver, vm);
} else {
virErrorPtr orig_err = virSaveLastError();
int reason;
/* cancel any outstanding NBD jobs */
qemuMigrationCancelDriveMirror(driver, vm, false,
QEMU_ASYNC_JOB_MIGRATION_OUT, NULL);
virSetError(orig_err);
virFreeError(orig_err);
if (virDomainObjGetState(vm, &reason) == VIR_DOMAIN_PAUSED &&
reason == VIR_DOMAIN_PAUSED_POSTCOPY) {
qemuMigrationPostcopyFailed(driver, vm);
} else if (qemuMigrationRestoreDomainState(conn, vm)) {
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_RESUMED,
VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
qemuDomainEventQueue(driver, event);
}
if (virDomainSaveStatus(driver->xmlopt, cfg->stateDir, vm, driver->caps) < 0)
VIR_WARN("Failed to save status on vm %s", vm->def->name);
}
done:
qemuMigrationCookieFree(mig);
rv = 0;
cleanup:
virObjectUnref(cfg);
return rv;
}
int
qemuMigrationConfirm(virConnectPtr conn,
virDomainObjPtr vm,
const char *cookiein,
int cookieinlen,
unsigned int flags,
int cancelled)
{
virQEMUDriverPtr driver = conn->privateData;
qemuMigrationJobPhase phase;
virQEMUDriverConfigPtr cfg = NULL;
int ret = -1;
cfg = virQEMUDriverGetConfig(driver);
if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_OUT))
goto cleanup;
if (cancelled)
phase = QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED;
else
phase = QEMU_MIGRATION_PHASE_CONFIRM3;
qemuMigrationJobStartPhase(driver, vm, phase);
virCloseCallbacksUnset(driver->closeCallbacks, vm,
qemuMigrationCleanup);
ret = qemuMigrationConfirmPhase(driver, conn, vm,
cookiein, cookieinlen,
flags, cancelled);
qemuMigrationJobFinish(driver, vm);
if (!virDomainObjIsActive(vm)) {
if (flags & VIR_MIGRATE_UNDEFINE_SOURCE) {
virDomainDeleteConfig(cfg->configDir, cfg->autostartDir, vm);
vm->persistent = 0;
}
qemuDomainRemoveInactive(driver, vm);
}
cleanup:
virDomainObjEndAPI(&vm);
virObjectUnref(cfg);
return ret;
}
enum qemuMigrationDestinationType {
MIGRATION_DEST_HOST,
MIGRATION_DEST_CONNECT_HOST,
MIGRATION_DEST_UNIX,
MIGRATION_DEST_FD,
};
enum qemuMigrationForwardType {
MIGRATION_FWD_DIRECT,
MIGRATION_FWD_STREAM,
};
typedef struct _qemuMigrationSpec qemuMigrationSpec;
typedef qemuMigrationSpec *qemuMigrationSpecPtr;
struct _qemuMigrationSpec {
enum qemuMigrationDestinationType destType;
union {
struct {
const char *protocol;
const char *name;
int port;
} host;
struct {
char *file;
int sock;
} unix_socket;
struct {
int qemu;
int local;
} fd;
} dest;
enum qemuMigrationForwardType fwdType;
union {
virStreamPtr stream;
} fwd;
};
#define TUNNEL_SEND_BUF_SIZE 65536
typedef struct _qemuMigrationIOThread qemuMigrationIOThread;
typedef qemuMigrationIOThread *qemuMigrationIOThreadPtr;
struct _qemuMigrationIOThread {
virThread thread;
virStreamPtr st;
int sock;
virError err;
int wakeupRecvFD;
int wakeupSendFD;
};
static void qemuMigrationIOFunc(void *arg)
{
qemuMigrationIOThreadPtr data = arg;
char *buffer = NULL;
struct pollfd fds[2];
int timeout = -1;
virErrorPtr err = NULL;
VIR_DEBUG("Running migration tunnel; stream=%p, sock=%d",
data->st, data->sock);
if (VIR_ALLOC_N(buffer, TUNNEL_SEND_BUF_SIZE) < 0)
goto abrt;
fds[0].fd = data->sock;
fds[1].fd = data->wakeupRecvFD;
for (;;) {
int ret;
fds[0].events = fds[1].events = POLLIN;
fds[0].revents = fds[1].revents = 0;
ret = poll(fds, ARRAY_CARDINALITY(fds), timeout);
if (ret < 0) {
if (errno == EAGAIN || errno == EINTR)
continue;
virReportSystemError(errno, "%s",
_("poll failed in migration tunnel"));
goto abrt;
}
if (ret == 0) {
/* We were asked to gracefully stop but reading would block. This
* can only happen if qemu told us migration finished but didn't
* close the migration fd. We handle this in the same way as EOF.
*/
VIR_DEBUG("QEMU forgot to close migration fd");
break;
}
if (fds[1].revents & (POLLIN | POLLERR | POLLHUP)) {
char stop = 0;
if (saferead(data->wakeupRecvFD, &stop, 1) != 1) {
virReportSystemError(errno, "%s",
_("failed to read from wakeup fd"));
goto abrt;
}
VIR_DEBUG("Migration tunnel was asked to %s",
stop ? "abort" : "finish");
if (stop) {
goto abrt;
} else {
timeout = 0;
}
}
if (fds[0].revents & (POLLIN | POLLERR | POLLHUP)) {
int nbytes;
nbytes = saferead(data->sock, buffer, TUNNEL_SEND_BUF_SIZE);
if (nbytes > 0) {
if (virStreamSend(data->st, buffer, nbytes) < 0)
goto error;
} else if (nbytes < 0) {
virReportSystemError(errno, "%s",
_("tunnelled migration failed to read from qemu"));
goto abrt;
} else {
/* EOF; get out of here */
break;
}
}
}
if (virStreamFinish(data->st) < 0)
goto error;
VIR_FORCE_CLOSE(data->sock);
VIR_FREE(buffer);
return;
abrt:
err = virSaveLastError();
if (err && err->code == VIR_ERR_OK) {
virFreeError(err);
err = NULL;
}
virStreamAbort(data->st);
if (err) {
virSetError(err);
virFreeError(err);
}
error:
/* Let the source qemu know that the transfer cant continue anymore.
* Don't copy the error for EPIPE as destination has the actual error. */
VIR_FORCE_CLOSE(data->sock);
if (!virLastErrorIsSystemErrno(EPIPE))
virCopyLastError(&data->err);
virResetLastError();
VIR_FREE(buffer);
}
static qemuMigrationIOThreadPtr
qemuMigrationStartTunnel(virStreamPtr st,
int sock)
{
qemuMigrationIOThreadPtr io = NULL;
int wakeupFD[2] = { -1, -1 };
if (pipe2(wakeupFD, O_CLOEXEC) < 0) {
virReportSystemError(errno, "%s",
_("Unable to make pipe"));
goto error;
}
if (VIR_ALLOC(io) < 0)
goto error;
io->st = st;
io->sock = sock;
io->wakeupRecvFD = wakeupFD[0];
io->wakeupSendFD = wakeupFD[1];
if (virThreadCreate(&io->thread, true,
qemuMigrationIOFunc,
io) < 0) {
virReportSystemError(errno, "%s",
_("Unable to create migration thread"));
goto error;
}
return io;
error:
VIR_FORCE_CLOSE(wakeupFD[0]);
VIR_FORCE_CLOSE(wakeupFD[1]);
VIR_FREE(io);
return NULL;
}
static int
qemuMigrationStopTunnel(qemuMigrationIOThreadPtr io, bool error)
{
int rv = -1;
char stop = error ? 1 : 0;
/* make sure the thread finishes its job and is joinable */
if (safewrite(io->wakeupSendFD, &stop, 1) != 1) {
virReportSystemError(errno, "%s",
_("failed to wakeup migration tunnel"));
goto cleanup;
}
virThreadJoin(&io->thread);
/* Forward error from the IO thread, to this thread */
if (io->err.code != VIR_ERR_OK) {
if (error)
rv = 0;
else
virSetError(&io->err);
virResetError(&io->err);
goto cleanup;
}
rv = 0;
cleanup:
VIR_FORCE_CLOSE(io->wakeupSendFD);
VIR_FORCE_CLOSE(io->wakeupRecvFD);
VIR_FREE(io);
return rv;
}
static int
qemuMigrationConnect(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMigrationSpecPtr spec)
{
virNetSocketPtr sock;
const char *host;
char *port = NULL;
int ret = -1;
host = spec->dest.host.name;
if (virAsprintf(&port, "%d", spec->dest.host.port) < 0)
return -1;
spec->destType = MIGRATION_DEST_FD;
spec->dest.fd.qemu = -1;
if (virSecurityManagerSetSocketLabel(driver->securityManager, vm->def) < 0)
goto cleanup;
if (virNetSocketNewConnectTCP(host, port,
AF_UNSPEC,
&sock) == 0) {
spec->dest.fd.qemu = virNetSocketDupFD(sock, true);
virObjectUnref(sock);
}
if (virSecurityManagerClearSocketLabel(driver->securityManager, vm->def) < 0 ||
spec->dest.fd.qemu == -1)
goto cleanup;
/* Migration expects a blocking FD */
if (virSetBlocking(spec->dest.fd.qemu, true) < 0) {
virReportSystemError(errno, _("Unable to set FD %d blocking"),
spec->dest.fd.qemu);
goto cleanup;
}
ret = 0;
cleanup:
VIR_FREE(port);
if (ret < 0)
VIR_FORCE_CLOSE(spec->dest.fd.qemu);
return ret;
}
static int
qemuMigrationRun(virQEMUDriverPtr driver,
virDomainObjPtr vm,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
unsigned long flags,
unsigned long resource,
qemuMigrationSpecPtr spec,
virConnectPtr dconn,
const char *graphicsuri,
size_t nmigrate_disks,
const char **migrate_disks)
{
int ret = -1;
unsigned int migrate_flags = QEMU_MONITOR_MIGRATE_BACKGROUND;
qemuDomainObjPrivatePtr priv = vm->privateData;
qemuMigrationCookiePtr mig = NULL;
qemuMigrationIOThreadPtr iothread = NULL;
int fd = -1;
unsigned long migrate_speed = resource ? resource : priv->migMaxBandwidth;
virErrorPtr orig_err = NULL;
unsigned int cookieFlags = 0;
bool abort_on_error = !!(flags & VIR_MIGRATE_ABORT_ON_ERROR);
bool events = virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATION_EVENT);
bool inPostCopy = false;
int rc;
VIR_DEBUG("driver=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
"cookieout=%p, cookieoutlen=%p, flags=%lx, resource=%lu, "
"spec=%p (dest=%d, fwd=%d), dconn=%p, graphicsuri=%s, "
"nmigrate_disks=%zu, migrate_disks=%p",
driver, vm, NULLSTR(cookiein), cookieinlen,
cookieout, cookieoutlen, flags, resource,
spec, spec->destType, spec->fwdType, dconn,
NULLSTR(graphicsuri), nmigrate_disks, migrate_disks);
if (flags & VIR_MIGRATE_NON_SHARED_DISK) {
migrate_flags |= QEMU_MONITOR_MIGRATE_NON_SHARED_DISK;
cookieFlags |= QEMU_MIGRATION_COOKIE_NBD;
}
if (flags & VIR_MIGRATE_NON_SHARED_INC) {
migrate_flags |= QEMU_MONITOR_MIGRATE_NON_SHARED_INC;
cookieFlags |= QEMU_MIGRATION_COOKIE_NBD;
}
if (virLockManagerPluginUsesState(driver->lockManager) &&
!cookieout) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Migration with lock driver %s requires"
" cookie support"),
virLockManagerPluginGetName(driver->lockManager));
return -1;
}
if (events)
priv->signalIOError = abort_on_error;
mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen,
cookieFlags | QEMU_MIGRATION_COOKIE_GRAPHICS);
if (!mig)
goto cleanup;
if (qemuDomainMigrateGraphicsRelocate(driver, vm, mig, graphicsuri) < 0)
VIR_WARN("unable to provide data for graphics client relocation");
if (migrate_flags & (QEMU_MONITOR_MIGRATE_NON_SHARED_DISK |
QEMU_MONITOR_MIGRATE_NON_SHARED_INC)) {
if (mig->nbd) {
/* This will update migrate_flags on success */
if (qemuMigrationDriveMirror(driver, vm, mig,
spec->dest.host.name,
migrate_speed,
&migrate_flags,
nmigrate_disks,
migrate_disks,
dconn) < 0) {
goto cleanup;
}
} else {
/* Destination doesn't support NBD server.
* Fall back to previous implementation. */
VIR_DEBUG("Destination doesn't support NBD server "
"Falling back to previous implementation.");
}
}
/* Before EnterMonitor, since qemuMigrationSetOffline already does that */
if (!(flags & VIR_MIGRATE_LIVE) &&
virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
if (qemuMigrationSetOffline(driver, vm) < 0)
goto cleanup;
}
if (qemuMigrationSetOption(driver, vm,
QEMU_MONITOR_MIGRATION_CAPS_XBZRLE,
flags & VIR_MIGRATE_COMPRESSED,
QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
if (qemuMigrationSetOption(driver, vm,
QEMU_MONITOR_MIGRATION_CAPS_AUTO_CONVERGE,
flags & VIR_MIGRATE_AUTO_CONVERGE,
QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
if (qemuMigrationSetOption(driver, vm,
QEMU_MONITOR_MIGRATION_CAPS_RDMA_PIN_ALL,
flags & VIR_MIGRATE_RDMA_PIN_ALL,
QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
if (qemuMigrationSetPostCopy(driver, vm,
flags & VIR_MIGRATE_POSTCOPY,
QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
if (qemuDomainObjEnterMonitorAsync(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
if (priv->job.abortJob) {
/* explicitly do this *after* we entered the monitor,
* as this is a critical section so we are guaranteed
* priv->job.abortJob will not change */
ignore_value(qemuDomainObjExitMonitor(driver, vm));
priv->job.current->type = VIR_DOMAIN_JOB_CANCELLED;
virReportError(VIR_ERR_OPERATION_ABORTED, _("%s: %s"),
qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
_("canceled by client"));
goto cleanup;
}
if (qemuMonitorSetMigrationSpeed(priv->mon, migrate_speed) < 0)
goto exit_monitor;
/* connect to the destination qemu if needed */
if (spec->destType == MIGRATION_DEST_CONNECT_HOST &&
qemuMigrationConnect(driver, vm, spec) < 0) {
goto exit_monitor;
}
switch (spec->destType) {
case MIGRATION_DEST_HOST:
if (STREQ(spec->dest.host.protocol, "rdma") &&
virProcessSetMaxMemLock(vm->pid, vm->def->mem.hard_limit << 10) < 0) {
goto exit_monitor;
}
ret = qemuMonitorMigrateToHost(priv->mon, migrate_flags,
spec->dest.host.protocol,
spec->dest.host.name,
spec->dest.host.port);
break;
case MIGRATION_DEST_CONNECT_HOST:
/* handled above and transformed into MIGRATION_DEST_FD */
break;
case MIGRATION_DEST_UNIX:
ret = qemuMonitorMigrateToUnix(priv->mon, migrate_flags,
spec->dest.unix_socket.file);
break;
case MIGRATION_DEST_FD:
if (spec->fwdType != MIGRATION_FWD_DIRECT) {
fd = spec->dest.fd.local;
spec->dest.fd.local = -1;
}
ret = qemuMonitorMigrateToFd(priv->mon, migrate_flags,
spec->dest.fd.qemu);
VIR_FORCE_CLOSE(spec->dest.fd.qemu);
break;
}
if (qemuDomainObjExitMonitor(driver, vm) < 0)
ret = -1;
if (ret < 0)
goto cleanup;
ret = -1;
/* From this point onwards we *must* call cancel to abort the
* migration on source if anything goes wrong */
if (spec->destType == MIGRATION_DEST_UNIX) {
/* It is also possible that the migrate didn't fail initially, but
* rather failed later on. Check its status before waiting for a
* connection from qemu which may never be initiated.
*/
if (qemuMigrationCheckJobStatus(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_OUT,
false) < 0)
goto cancel;
while ((fd = accept(spec->dest.unix_socket.sock, NULL, NULL)) < 0) {
if (errno == EAGAIN || errno == EINTR)
continue;
virReportSystemError(errno, "%s",
_("failed to accept connection from qemu"));
goto cancel;
}
}
if (spec->fwdType != MIGRATION_FWD_DIRECT) {
if (!(iothread = qemuMigrationStartTunnel(spec->fwd.stream, fd)))
goto cancel;
/* If we've created a tunnel, then the 'fd' will be closed in the
* qemuMigrationIOFunc as data->sock.
*/
fd = -1;
}
rc = qemuMigrationWaitForCompletion(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_OUT,
dconn, abort_on_error, !!mig->nbd);
if (rc == -2)
goto cancel;
else if (rc == -1)
goto cleanup;
if (priv->job.current->stats.status == QEMU_MONITOR_MIGRATION_STATUS_POSTCOPY)
inPostCopy = true;
/* When migration completed, QEMU will have paused the CPUs for us.
* Wait for the STOP event to be processed or explicitly stop CPUs
* (for old QEMU which does not send events) to release the lock state.
*/
if (priv->monJSON) {
while (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
priv->signalStop = true;
rc = virDomainObjWait(vm);
priv->signalStop = false;
if (rc < 0)
goto cancelPostCopy;
}
} else if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING &&
qemuMigrationSetOffline(driver, vm) < 0) {
goto cancelPostCopy;
}
if (priv->job.completed)
priv->job.completed->stopped = priv->job.current->stopped;
ret = 0;
cleanup:
if (ret < 0 && !orig_err)
orig_err = virSaveLastError();
/* cancel any outstanding NBD jobs */
if (mig && mig->nbd) {
if (qemuMigrationCancelDriveMirror(driver, vm, ret == 0,
QEMU_ASYNC_JOB_MIGRATION_OUT,
dconn) < 0)
ret = -1;
}
if (spec->fwdType != MIGRATION_FWD_DIRECT) {
if (iothread && qemuMigrationStopTunnel(iothread, ret < 0) < 0)
ret = -1;
}
VIR_FORCE_CLOSE(fd);
if (priv->job.completed) {
qemuDomainJobInfoUpdateTime(priv->job.completed);
qemuDomainJobInfoUpdateDowntime(priv->job.completed);
ignore_value(virTimeMillisNow(&priv->job.completed->sent));
}
if (priv->job.current->type == VIR_DOMAIN_JOB_UNBOUNDED && !inPostCopy)
priv->job.current->type = VIR_DOMAIN_JOB_FAILED;
cookieFlags |= QEMU_MIGRATION_COOKIE_NETWORK |
QEMU_MIGRATION_COOKIE_STATS;
if (flags & VIR_MIGRATE_PERSIST_DEST)
cookieFlags |= QEMU_MIGRATION_COOKIE_PERSISTENT;
if (ret == 0 &&
qemuMigrationBakeCookie(mig, driver, vm, cookieout,
cookieoutlen, cookieFlags) < 0) {
VIR_WARN("Unable to encode migration cookie");
}
qemuMigrationCookieFree(mig);
if (events)
priv->signalIOError = false;
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
return ret;
exit_monitor:
ignore_value(qemuDomainObjExitMonitor(driver, vm));
goto cleanup;
cancel:
orig_err = virSaveLastError();
if (virDomainObjIsActive(vm)) {
if (qemuDomainObjEnterMonitorAsync(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_OUT) == 0) {
qemuMonitorMigrateCancel(priv->mon);
ignore_value(qemuDomainObjExitMonitor(driver, vm));
}
}
goto cleanup;
cancelPostCopy:
priv->job.current->type = VIR_DOMAIN_JOB_FAILED;
if (inPostCopy)
goto cancel;
else
goto cleanup;
}
/* Perform migration using QEMU's native migrate support,
* not encrypted obviously
*/
static int doNativeMigrate(virQEMUDriverPtr driver,
virDomainObjPtr vm,
const char *uri,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
unsigned long flags,
unsigned long resource,
virConnectPtr dconn,
const char *graphicsuri,
size_t nmigrate_disks,
const char **migrate_disks)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
virURIPtr uribits = NULL;
int ret = -1;
qemuMigrationSpec spec;
VIR_DEBUG("driver=%p, vm=%p, uri=%s, cookiein=%s, cookieinlen=%d, "
"cookieout=%p, cookieoutlen=%p, flags=%lx, resource=%lu, "
"graphicsuri=%s, nmigrate_disks=%zu migrate_disks=%p",
driver, vm, uri, NULLSTR(cookiein), cookieinlen,
cookieout, cookieoutlen, flags, resource,
NULLSTR(graphicsuri), nmigrate_disks, migrate_disks);
if (!(uribits = qemuMigrationParseURI(uri, NULL)))
return -1;
if (uribits->scheme == NULL) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("missing scheme in migration URI: %s"),
uri);
goto cleanup;
}
if (STREQ(uribits->scheme, "rdma")) {
if (!virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_RDMA)) {
virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s",
_("outgoing RDMA migration is not supported "
"with this QEMU binary"));
goto cleanup;
}
if (!virMemoryLimitIsSet(vm->def->mem.hard_limit)) {
virReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("cannot start RDMA migration with no memory hard "
"limit set"));
goto cleanup;
}
}
if (STRNEQ(uribits->scheme, "rdma"))
spec.destType = MIGRATION_DEST_CONNECT_HOST;
else
spec.destType = MIGRATION_DEST_HOST;
spec.dest.host.protocol = uribits->scheme;
spec.dest.host.name = uribits->server;
spec.dest.host.port = uribits->port;
spec.fwdType = MIGRATION_FWD_DIRECT;
ret = qemuMigrationRun(driver, vm, cookiein, cookieinlen, cookieout,
cookieoutlen, flags, resource, &spec, dconn,
graphicsuri, nmigrate_disks, migrate_disks);
if (spec.destType == MIGRATION_DEST_FD)
VIR_FORCE_CLOSE(spec.dest.fd.qemu);
cleanup:
virURIFree(uribits);
return ret;
}
static int doTunnelMigrate(virQEMUDriverPtr driver,
virDomainObjPtr vm,
virStreamPtr st,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
unsigned long flags,
unsigned long resource,
virConnectPtr dconn,
const char *graphicsuri,
size_t nmigrate_disks,
const char **migrate_disks)
{
virNetSocketPtr sock = NULL;
int ret = -1;
qemuMigrationSpec spec;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
int fds[2] = { -1, -1 };
VIR_DEBUG("driver=%p, vm=%p, st=%p, cookiein=%s, cookieinlen=%d, "
"cookieout=%p, cookieoutlen=%p, flags=%lx, resource=%lu, "
"graphicsuri=%s, nmigrate_disks=%zu, migrate_disks=%p",
driver, vm, st, NULLSTR(cookiein), cookieinlen,
cookieout, cookieoutlen, flags, resource,
NULLSTR(graphicsuri), nmigrate_disks, migrate_disks);
spec.fwdType = MIGRATION_FWD_STREAM;
spec.fwd.stream = st;
spec.destType = MIGRATION_DEST_FD;
spec.dest.fd.qemu = -1;
spec.dest.fd.local = -1;
if (pipe2(fds, O_CLOEXEC) == 0) {
spec.dest.fd.qemu = fds[1];
spec.dest.fd.local = fds[0];
}
if (spec.dest.fd.qemu == -1 ||
virSecurityManagerSetImageFDLabel(driver->securityManager, vm->def,
spec.dest.fd.qemu) < 0) {
virReportSystemError(errno, "%s",
_("cannot create pipe for tunnelled migration"));
goto cleanup;
}
ret = qemuMigrationRun(driver, vm, cookiein, cookieinlen, cookieout,
cookieoutlen, flags, resource, &spec, dconn,
graphicsuri, nmigrate_disks, migrate_disks);
cleanup:
if (spec.destType == MIGRATION_DEST_FD) {
VIR_FORCE_CLOSE(spec.dest.fd.qemu);
VIR_FORCE_CLOSE(spec.dest.fd.local);
} else {
virObjectUnref(sock);
VIR_FREE(spec.dest.unix_socket.file);
}
virObjectUnref(cfg);
return ret;
}
/* This is essentially a re-impl of virDomainMigrateVersion2
* from libvirt.c, but running in source libvirtd context,
* instead of client app context & also adding in tunnel
* handling */
static int doPeer2PeerMigrate2(virQEMUDriverPtr driver,
virConnectPtr sconn ATTRIBUTE_UNUSED,
virConnectPtr dconn,
virDomainObjPtr vm,
const char *dconnuri,
unsigned long flags,
const char *dname,
unsigned long resource)
{
virDomainPtr ddomain = NULL;
char *uri_out = NULL;
char *cookie = NULL;
char *dom_xml = NULL;
int cookielen = 0, ret;
virErrorPtr orig_err = NULL;
bool cancelled;
virStreamPtr st = NULL;
unsigned long destflags;
VIR_DEBUG("driver=%p, sconn=%p, dconn=%p, vm=%p, dconnuri=%s, "
"flags=%lx, dname=%s, resource=%lu",
driver, sconn, dconn, vm, NULLSTR(dconnuri),
flags, NULLSTR(dname), resource);
/* In version 2 of the protocol, the prepare step is slightly
* different. We fetch the domain XML of the source domain
* and pass it to Prepare2.
*/
if (!(dom_xml = qemuDomainFormatXML(driver, vm,
QEMU_DOMAIN_FORMAT_LIVE_FLAGS |
VIR_DOMAIN_XML_MIGRATABLE)))
return -1;
if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED)
flags |= VIR_MIGRATE_PAUSED;
destflags = flags & ~(VIR_MIGRATE_ABORT_ON_ERROR |
VIR_MIGRATE_AUTO_CONVERGE);
VIR_DEBUG("Prepare2 %p", dconn);
if (flags & VIR_MIGRATE_TUNNELLED) {
/*
* Tunnelled Migrate Version 2 does not support cookies
* due to missing parameters in the prepareTunnel() API.
*/
if (!(st = virStreamNew(dconn, 0)))
goto cleanup;
qemuDomainObjEnterRemote(vm);
ret = dconn->driver->domainMigratePrepareTunnel
(dconn, st, destflags, dname, resource, dom_xml);
qemuDomainObjExitRemote(vm);
} else {
qemuDomainObjEnterRemote(vm);
ret = dconn->driver->domainMigratePrepare2
(dconn, &cookie, &cookielen, NULL, &uri_out,
destflags, dname, resource, dom_xml);
qemuDomainObjExitRemote(vm);
}
VIR_FREE(dom_xml);
if (ret == -1)
goto cleanup;
/* the domain may have shutdown or crashed while we had the locks dropped
* in qemuDomainObjEnterRemote, so check again
*/
if (!virDomainObjIsActive(vm)) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("guest unexpectedly quit"));
goto cleanup;
}
if (!(flags & VIR_MIGRATE_TUNNELLED) &&
(uri_out == NULL)) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("domainMigratePrepare2 did not set uri"));
cancelled = true;
orig_err = virSaveLastError();
goto finish;
}
/* Perform the migration. The driver isn't supposed to return
* until the migration is complete.
*/
VIR_DEBUG("Perform %p", sconn);
qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM2);
if (flags & VIR_MIGRATE_TUNNELLED)
ret = doTunnelMigrate(driver, vm, st,
NULL, 0, NULL, NULL,
flags, resource, dconn,
NULL, 0, NULL);
else
ret = doNativeMigrate(driver, vm, uri_out,
cookie, cookielen,
NULL, NULL, /* No out cookie with v2 migration */
flags, resource, dconn, NULL, 0, NULL);
/* Perform failed. Make sure Finish doesn't overwrite the error */
if (ret < 0)
orig_err = virSaveLastError();
/* If Perform returns < 0, then we need to cancel the VM
* startup on the destination
*/
cancelled = ret < 0;
finish:
/* In version 2 of the migration protocol, we pass the
* status code from the sender to the destination host,
* so it can do any cleanup if the migration failed.
*/
dname = dname ? dname : vm->def->name;
VIR_DEBUG("Finish2 %p ret=%d", dconn, ret);
qemuDomainObjEnterRemote(vm);
ddomain = dconn->driver->domainMigrateFinish2
(dconn, dname, cookie, cookielen,
uri_out ? uri_out : dconnuri, destflags, cancelled);
qemuDomainObjExitRemote(vm);
if (cancelled && ddomain)
VIR_ERROR(_("finish step ignored that migration was cancelled"));
cleanup:
if (ddomain) {
virObjectUnref(ddomain);
ret = 0;
} else {
ret = -1;
}
virObjectUnref(st);
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
VIR_FREE(uri_out);
VIR_FREE(cookie);
return ret;
}
/* This is essentially a re-impl of virDomainMigrateVersion3
* from libvirt.c, but running in source libvirtd context,
* instead of client app context & also adding in tunnel
* handling */
static int
doPeer2PeerMigrate3(virQEMUDriverPtr driver,
virConnectPtr sconn,
virConnectPtr dconn,
const char *dconnuri,
virDomainObjPtr vm,
const char *xmlin,
const char *dname,
const char *uri,
const char *graphicsuri,
const char *listenAddress,
size_t nmigrate_disks,
const char **migrate_disks,
int nbdPort,
unsigned long long bandwidth,
bool useParams,
unsigned long flags)
{
virDomainPtr ddomain = NULL;
char *uri_out = NULL;
char *cookiein = NULL;
char *cookieout = NULL;
char *dom_xml = NULL;
int cookieinlen = 0;
int cookieoutlen = 0;
int ret = -1;
virErrorPtr orig_err = NULL;
bool cancelled = true;
virStreamPtr st = NULL;
unsigned long destflags;
virTypedParameterPtr params = NULL;
int nparams = 0;
int maxparams = 0;
size_t i;
VIR_DEBUG("driver=%p, sconn=%p, dconn=%p, dconnuri=%s, vm=%p, xmlin=%s, "
"dname=%s, uri=%s, graphicsuri=%s, listenAddress=%s, "
"nmigrate_disks=%zu, migrate_disks=%p, nbdPort=%d, "
"bandwidth=%llu, useParams=%d, flags=%lx",
driver, sconn, dconn, NULLSTR(dconnuri), vm, NULLSTR(xmlin),
NULLSTR(dname), NULLSTR(uri), NULLSTR(graphicsuri),
NULLSTR(listenAddress), nmigrate_disks, migrate_disks, nbdPort,
bandwidth, useParams, flags);
/* Unlike the virDomainMigrateVersion3 counterpart, we don't need
* to worry about auto-setting the VIR_MIGRATE_CHANGE_PROTECTION
* bit here, because we are already running inside the context of
* a single job. */
dom_xml = qemuMigrationBeginPhase(driver, vm, xmlin, dname,
&cookieout, &cookieoutlen,
nmigrate_disks, migrate_disks, flags);
if (!dom_xml)
goto cleanup;
if (useParams) {
if (virTypedParamsAddString(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_DEST_XML, dom_xml) < 0)
goto cleanup;
if (dname &&
virTypedParamsAddString(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_DEST_NAME, dname) < 0)
goto cleanup;
if (uri &&
virTypedParamsAddString(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_URI, uri) < 0)
goto cleanup;
if (bandwidth &&
virTypedParamsAddULLong(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_BANDWIDTH,
bandwidth) < 0)
goto cleanup;
if (graphicsuri &&
virTypedParamsAddString(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_GRAPHICS_URI,
graphicsuri) < 0)
goto cleanup;
if (listenAddress &&
virTypedParamsAddString(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_LISTEN_ADDRESS,
listenAddress) < 0)
goto cleanup;
for (i = 0; i < nmigrate_disks; i++)
if (virTypedParamsAddString(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_MIGRATE_DISKS,
migrate_disks[i]) < 0)
goto cleanup;
if (nbdPort &&
virTypedParamsAddInt(¶ms, &nparams, &maxparams,
VIR_MIGRATE_PARAM_DISKS_PORT,
nbdPort) < 0)
goto cleanup;
}
if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED)
flags |= VIR_MIGRATE_PAUSED;
destflags = flags & ~(VIR_MIGRATE_ABORT_ON_ERROR |
VIR_MIGRATE_AUTO_CONVERGE);
VIR_DEBUG("Prepare3 %p", dconn);
cookiein = cookieout;
cookieinlen = cookieoutlen;
cookieout = NULL;
cookieoutlen = 0;
if (flags & VIR_MIGRATE_TUNNELLED) {
if (!(st = virStreamNew(dconn, 0)))
goto cleanup;
qemuDomainObjEnterRemote(vm);
if (useParams) {
ret = dconn->driver->domainMigratePrepareTunnel3Params
(dconn, st, params, nparams, cookiein, cookieinlen,
&cookieout, &cookieoutlen, destflags);
} else {
ret = dconn->driver->domainMigratePrepareTunnel3
(dconn, st, cookiein, cookieinlen, &cookieout, &cookieoutlen,
destflags, dname, bandwidth, dom_xml);
}
qemuDomainObjExitRemote(vm);
} else {
qemuDomainObjEnterRemote(vm);
if (useParams) {
ret = dconn->driver->domainMigratePrepare3Params
(dconn, params, nparams, cookiein, cookieinlen,
&cookieout, &cookieoutlen, &uri_out, destflags);
} else {
ret = dconn->driver->domainMigratePrepare3
(dconn, cookiein, cookieinlen, &cookieout, &cookieoutlen,
uri, &uri_out, destflags, dname, bandwidth, dom_xml);
}
qemuDomainObjExitRemote(vm);
}
VIR_FREE(dom_xml);
if (ret == -1)
goto cleanup;
if (flags & VIR_MIGRATE_OFFLINE) {
VIR_DEBUG("Offline migration, skipping Perform phase");
VIR_FREE(cookieout);
cookieoutlen = 0;
cancelled = false;
goto finish;
}
if (uri_out) {
uri = uri_out;
if (useParams &&
virTypedParamsReplaceString(¶ms, &nparams,
VIR_MIGRATE_PARAM_URI, uri_out) < 0) {
orig_err = virSaveLastError();
goto finish;
}
} else if (!uri && !(flags & VIR_MIGRATE_TUNNELLED)) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("domainMigratePrepare3 did not set uri"));
orig_err = virSaveLastError();
goto finish;
}
/* Perform the migration. The driver isn't supposed to return
* until the migration is complete. The src VM should remain
* running, but in paused state until the destination can
* confirm migration completion.
*/
VIR_DEBUG("Perform3 %p uri=%s", sconn, NULLSTR(uri));
qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM3);
VIR_FREE(cookiein);
cookiein = cookieout;
cookieinlen = cookieoutlen;
cookieout = NULL;
cookieoutlen = 0;
if (flags & VIR_MIGRATE_TUNNELLED) {
ret = doTunnelMigrate(driver, vm, st,
cookiein, cookieinlen,
&cookieout, &cookieoutlen,
flags, bandwidth, dconn, graphicsuri,
nmigrate_disks, migrate_disks);
} else {
ret = doNativeMigrate(driver, vm, uri,
cookiein, cookieinlen,
&cookieout, &cookieoutlen,
flags, bandwidth, dconn, graphicsuri,
nmigrate_disks, migrate_disks);
}
/* Perform failed. Make sure Finish doesn't overwrite the error */
if (ret < 0) {
orig_err = virSaveLastError();
} else {
qemuMigrationJobSetPhase(driver, vm,
QEMU_MIGRATION_PHASE_PERFORM3_DONE);
}
/* If Perform returns < 0, then we need to cancel the VM
* startup on the destination
*/
cancelled = ret < 0;
finish:
/*
* The status code from the source is passed to the destination.
* The dest can cleanup in the source indicated it failed to
* send all migration data. Returns NULL for ddomain if
* the dest was unable to complete migration.
*/
VIR_DEBUG("Finish3 %p ret=%d", dconn, ret);
VIR_FREE(cookiein);
cookiein = cookieout;
cookieinlen = cookieoutlen;
cookieout = NULL;
cookieoutlen = 0;
if (useParams) {
if (virTypedParamsGetString(params, nparams,
VIR_MIGRATE_PARAM_DEST_NAME, NULL) <= 0 &&
virTypedParamsReplaceString(¶ms, &nparams,
VIR_MIGRATE_PARAM_DEST_NAME,
vm->def->name) < 0) {
ddomain = NULL;
} else {
qemuDomainObjEnterRemote(vm);
ddomain = dconn->driver->domainMigrateFinish3Params
(dconn, params, nparams, cookiein, cookieinlen,
&cookieout, &cookieoutlen, destflags, cancelled);
qemuDomainObjExitRemote(vm);
}
} else {
dname = dname ? dname : vm->def->name;
qemuDomainObjEnterRemote(vm);
ddomain = dconn->driver->domainMigrateFinish3
(dconn, dname, cookiein, cookieinlen, &cookieout, &cookieoutlen,
dconnuri, uri, destflags, cancelled);
qemuDomainObjExitRemote(vm);
}
if (cancelled) {
if (ddomain) {
VIR_ERROR(_("finish step ignored that migration was cancelled"));
} else {
/* If Finish reported a useful error, use it instead of the
* original "migration unexpectedly failed" error.
*
* This is ugly but we can't do better with the APIs we have. We
* only replace the error if Finish was called with cancelled == 1
* and reported a real error (old libvirt would report an error
* from RPC instead of MIGRATE_FINISH_OK), which only happens when
* the domain died on destination. To further reduce a possibility
* of false positives we also check that Perform returned
* VIR_ERR_OPERATION_FAILED.
*/
if (orig_err &&
orig_err->domain == VIR_FROM_QEMU &&
orig_err->code == VIR_ERR_OPERATION_FAILED) {
virErrorPtr err = virGetLastError();
if (err &&
err->domain == VIR_FROM_QEMU &&
err->code != VIR_ERR_MIGRATE_FINISH_OK) {
virFreeError(orig_err);
orig_err = NULL;
}
}
}
}
/* If ddomain is NULL, then we were unable to start
* the guest on the target, and must restart on the
* source. There is a small chance that the ddomain
* is NULL due to an RPC failure, in which case
* ddomain could in fact be running on the dest.
* The lock manager plugins should take care of
* safety in this scenario.
*/
cancelled = ddomain == NULL;
/* If finish3 set an error, and we don't have an earlier
* one we need to preserve it in case confirm3 overwrites
*/
if (!orig_err)
orig_err = virSaveLastError();
/*
* If cancelled, then src VM will be restarted, else
* it will be killed
*/
VIR_DEBUG("Confirm3 %p cancelled=%d vm=%p", sconn, cancelled, vm);
VIR_FREE(cookiein);
cookiein = cookieout;
cookieinlen = cookieoutlen;
cookieout = NULL;
cookieoutlen = 0;
ret = qemuMigrationConfirmPhase(driver, sconn, vm,
cookiein, cookieinlen,
flags, cancelled);
/* If Confirm3 returns -1, there's nothing more we can
* do, but fortunately worst case is that there is a
* domain left in 'paused' state on source.
*/
if (ret < 0)
VIR_WARN("Guest %s probably left in 'paused' state on source",
vm->def->name);
cleanup:
if (ddomain) {
virObjectUnref(ddomain);
ret = 0;
} else {
ret = -1;
}
virObjectUnref(st);
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
VIR_FREE(uri_out);
VIR_FREE(cookiein);
VIR_FREE(cookieout);
virTypedParamsFree(params, nparams);
return ret;
}
static void
qemuMigrationConnectionClosed(virConnectPtr conn,
int reason,
void *opaque)
{
virDomainObjPtr vm = opaque;
VIR_DEBUG("conn=%p, reason=%d, vm=%s", conn, reason, vm->def->name);
virDomainObjBroadcast(vm);
}
static int virConnectCredType[] = {
VIR_CRED_AUTHNAME,
VIR_CRED_PASSPHRASE,
};
static virConnectAuth virConnectAuthConfig = {
.credtype = virConnectCredType,
.ncredtype = ARRAY_CARDINALITY(virConnectCredType),
};
static int doPeer2PeerMigrate(virQEMUDriverPtr driver,
virConnectPtr sconn,
virDomainObjPtr vm,
const char *xmlin,
const char *dconnuri,
const char *uri,
const char *graphicsuri,
const char *listenAddress,
size_t nmigrate_disks,
const char **migrate_disks,
int nbdPort,
unsigned long flags,
const char *dname,
unsigned long resource,
bool *v3proto)
{
int ret = -1;
virConnectPtr dconn = NULL;
bool p2p;
virErrorPtr orig_err = NULL;
bool offline = false;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
bool useParams;
VIR_DEBUG("driver=%p, sconn=%p, vm=%p, xmlin=%s, dconnuri=%s, uri=%s, "
"graphicsuri=%s, listenAddress=%s, nmigrate_disks=%zu, "
"migrate_disks=%p, nbdPort=%d, flags=%lx, dname=%s, "
"resource=%lu",
driver, sconn, vm, NULLSTR(xmlin), NULLSTR(dconnuri),
NULLSTR(uri), NULLSTR(graphicsuri), NULLSTR(listenAddress),
nmigrate_disks, migrate_disks, nbdPort, flags, NULLSTR(dname),
resource);
if (flags & VIR_MIGRATE_TUNNELLED && uri) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("migration URI is not supported by tunnelled "
"migration"));
goto cleanup;
}
if (flags & VIR_MIGRATE_TUNNELLED && listenAddress) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("listen address is not supported by tunnelled "
"migration"));
goto cleanup;
}
if (flags & VIR_MIGRATE_TUNNELLED && nbdPort) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("disk port address is not supported by tunnelled "
"migration"));
goto cleanup;
}
/* the order of operations is important here; we make sure the
* destination side is completely setup before we touch the source
*/
qemuDomainObjEnterRemote(vm);
dconn = virConnectOpenAuth(dconnuri, &virConnectAuthConfig, 0);
qemuDomainObjExitRemote(vm);
if (dconn == NULL) {
virReportError(VIR_ERR_OPERATION_FAILED,
_("Failed to connect to remote libvirt URI %s: %s"),
dconnuri, virGetLastErrorMessage());
virObjectUnref(cfg);
return -1;
}
if (virConnectSetKeepAlive(dconn, cfg->keepAliveInterval,
cfg->keepAliveCount) < 0)
goto cleanup;
if (virConnectRegisterCloseCallback(dconn, qemuMigrationConnectionClosed,
vm, NULL) < 0) {
goto cleanup;
}
qemuDomainObjEnterRemote(vm);
p2p = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
VIR_DRV_FEATURE_MIGRATION_P2P);
/* v3proto reflects whether the caller used Perform3, but with
* p2p migrate, regardless of whether Perform2 or Perform3
* were used, we decide protocol based on what target supports
*/
*v3proto = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
VIR_DRV_FEATURE_MIGRATION_V3);
useParams = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
VIR_DRV_FEATURE_MIGRATION_PARAMS);
if (flags & VIR_MIGRATE_OFFLINE)
offline = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
VIR_DRV_FEATURE_MIGRATION_OFFLINE);
qemuDomainObjExitRemote(vm);
if (!p2p) {
virReportError(VIR_ERR_OPERATION_FAILED, "%s",
_("Destination libvirt does not support peer-to-peer migration protocol"));
goto cleanup;
}
/* Only xmlin, dname, uri, and bandwidth parameters can be used with
* old-style APIs. */
if (!useParams && (graphicsuri || listenAddress || nmigrate_disks)) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("Migration APIs with extensible parameters are not "
"supported but extended parameters were passed"));
goto cleanup;
}
if (flags & VIR_MIGRATE_OFFLINE && !offline) {
virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
_("offline migration is not supported by "
"the destination host"));
goto cleanup;
}
/* domain may have been stopped while we were talking to remote daemon */
if (!virDomainObjIsActive(vm) && !(flags & VIR_MIGRATE_OFFLINE)) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("guest unexpectedly quit"));
goto cleanup;
}
/* Change protection is only required on the source side (us), and
* only for v3 migration when begin and perform are separate jobs.
* But peer-2-peer is already a single job, and we still want to
* talk to older destinations that would reject the flag.
* Therefore it is safe to clear the bit here. */
flags &= ~VIR_MIGRATE_CHANGE_PROTECTION;
if (*v3proto) {
ret = doPeer2PeerMigrate3(driver, sconn, dconn, dconnuri, vm, xmlin,
dname, uri, graphicsuri, listenAddress,
nmigrate_disks, migrate_disks, nbdPort,
resource, useParams, flags);
} else {
ret = doPeer2PeerMigrate2(driver, sconn, dconn, vm,
dconnuri, flags, dname, resource);
}
cleanup:
orig_err = virSaveLastError();
qemuDomainObjEnterRemote(vm);
virConnectUnregisterCloseCallback(dconn, qemuMigrationConnectionClosed);
virObjectUnref(dconn);
qemuDomainObjExitRemote(vm);
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
virObjectUnref(cfg);
return ret;
}
/*
* This implements perform part of the migration protocol when migration job
* does not need to be active across several APIs, i.e., peer2peer migration or
* perform phase of v2 non-peer2peer migration.
*/
static int
qemuMigrationPerformJob(virQEMUDriverPtr driver,
virConnectPtr conn,
virDomainObjPtr vm,
const char *xmlin,
const char *dconnuri,
const char *uri,
const char *graphicsuri,
const char *listenAddress,
size_t nmigrate_disks,
const char **migrate_disks,
int nbdPort,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
unsigned long flags,
const char *dname,
unsigned long resource,
bool v3proto)
{
virObjectEventPtr event = NULL;
int ret = -1;
virErrorPtr orig_err = NULL;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
if (!virDomainObjIsActive(vm) && !(flags & VIR_MIGRATE_OFFLINE)) {
virReportError(VIR_ERR_OPERATION_INVALID,
"%s", _("domain is not running"));
goto endjob;
}
if (!qemuMigrationIsAllowed(driver, vm, true, flags))
goto endjob;
if (!(flags & VIR_MIGRATE_UNSAFE) &&
!qemuMigrationIsSafe(vm->def, nmigrate_disks, migrate_disks))
goto endjob;
qemuMigrationStoreDomainState(vm);
if ((flags & (VIR_MIGRATE_TUNNELLED | VIR_MIGRATE_PEER2PEER))) {
ret = doPeer2PeerMigrate(driver, conn, vm, xmlin,
dconnuri, uri, graphicsuri, listenAddress,
nmigrate_disks, migrate_disks, nbdPort,
flags, dname, resource, &v3proto);
} else {
qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM2);
ret = doNativeMigrate(driver, vm, uri, cookiein, cookieinlen,
cookieout, cookieoutlen,
flags, resource, NULL, NULL, 0, NULL);
}
if (ret < 0)
goto endjob;
/*
* In v3 protocol, the source VM is not killed off until the
* confirm step.
*/
if (!v3proto) {
qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_MIGRATED,
QEMU_ASYNC_JOB_MIGRATION_OUT,
VIR_QEMU_PROCESS_STOP_MIGRATED);
virDomainAuditStop(vm, "migrated");
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_STOPPED,
VIR_DOMAIN_EVENT_STOPPED_MIGRATED);
}
endjob:
if (ret < 0)
orig_err = virSaveLastError();
if (qemuMigrationRestoreDomainState(conn, vm)) {
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_RESUMED,
VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
}
qemuMigrationJobFinish(driver, vm);
if (!virDomainObjIsActive(vm) && ret == 0) {
if (flags & VIR_MIGRATE_UNDEFINE_SOURCE) {
virDomainDeleteConfig(cfg->configDir, cfg->autostartDir, vm);
vm->persistent = 0;
}
qemuDomainRemoveInactive(driver, vm);
}
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
cleanup:
virDomainObjEndAPI(&vm);
qemuDomainEventQueue(driver, event);
virObjectUnref(cfg);
return ret;
}
/*
* This implements perform phase of v3 migration protocol.
*/
static int
qemuMigrationPerformPhase(virQEMUDriverPtr driver,
virConnectPtr conn,
virDomainObjPtr vm,
const char *uri,
const char *graphicsuri,
size_t nmigrate_disks,
const char **migrate_disks,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
unsigned long flags,
unsigned long resource)
{
virObjectEventPtr event = NULL;
int ret = -1;
/* If we didn't start the job in the begin phase, start it now. */
if (!(flags & VIR_MIGRATE_CHANGE_PROTECTION)) {
if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
goto cleanup;
} else if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_OUT)) {
goto cleanup;
}
qemuMigrationJobStartPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM3);
virCloseCallbacksUnset(driver->closeCallbacks, vm,
qemuMigrationCleanup);
ret = doNativeMigrate(driver, vm, uri, cookiein, cookieinlen,
cookieout, cookieoutlen,
flags, resource, NULL, graphicsuri,
nmigrate_disks, migrate_disks);
if (ret < 0) {
if (qemuMigrationRestoreDomainState(conn, vm)) {
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_RESUMED,
VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
}
goto endjob;
}
qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM3_DONE);
if (virCloseCallbacksSet(driver->closeCallbacks, vm, conn,
qemuMigrationCleanup) < 0)
goto endjob;
endjob:
if (ret < 0)
qemuMigrationJobFinish(driver, vm);
else
qemuMigrationJobContinue(vm);
if (!virDomainObjIsActive(vm))
qemuDomainRemoveInactive(driver, vm);
cleanup:
virDomainObjEndAPI(&vm);
qemuDomainEventQueue(driver, event);
return ret;
}
int
qemuMigrationPerform(virQEMUDriverPtr driver,
virConnectPtr conn,
virDomainObjPtr vm,
const char *xmlin,
const char *dconnuri,
const char *uri,
const char *graphicsuri,
const char *listenAddress,
size_t nmigrate_disks,
const char **migrate_disks,
int nbdPort,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
unsigned long flags,
const char *dname,
unsigned long resource,
bool v3proto)
{
VIR_DEBUG("driver=%p, conn=%p, vm=%p, xmlin=%s, dconnuri=%s, "
"uri=%s, graphicsuri=%s, listenAddress=%s, "
"nmigrate_disks=%zu, migrate_disks=%p, nbdPort=%d, "
"cookiein=%s, cookieinlen=%d, cookieout=%p, cookieoutlen=%p, "
"flags=%lx, dname=%s, resource=%lu, v3proto=%d",
driver, conn, vm, NULLSTR(xmlin), NULLSTR(dconnuri),
NULLSTR(uri), NULLSTR(graphicsuri), NULLSTR(listenAddress),
nmigrate_disks, migrate_disks, nbdPort,
NULLSTR(cookiein), cookieinlen, cookieout, cookieoutlen,
flags, NULLSTR(dname), resource, v3proto);
if ((flags & (VIR_MIGRATE_TUNNELLED | VIR_MIGRATE_PEER2PEER))) {
if (cookieinlen) {
virReportError(VIR_ERR_OPERATION_INVALID,
"%s", _("received unexpected cookie with P2P migration"));
return -1;
}
return qemuMigrationPerformJob(driver, conn, vm, xmlin, dconnuri, uri,
graphicsuri, listenAddress,
nmigrate_disks, migrate_disks, nbdPort,
cookiein, cookieinlen,
cookieout, cookieoutlen,
flags, dname, resource, v3proto);
} else {
if (dconnuri) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("Unexpected dconnuri parameter with non-peer2peer migration"));
return -1;
}
if (v3proto) {
return qemuMigrationPerformPhase(driver, conn, vm, uri,
graphicsuri,
nmigrate_disks, migrate_disks,
cookiein, cookieinlen,
cookieout, cookieoutlen,
flags, resource);
} else {
return qemuMigrationPerformJob(driver, conn, vm, xmlin, NULL,
uri, graphicsuri, listenAddress,
nmigrate_disks, migrate_disks, nbdPort,
cookiein, cookieinlen,
cookieout, cookieoutlen, flags,
dname, resource, v3proto);
}
}
}
static int
qemuMigrationVPAssociatePortProfiles(virDomainDefPtr def)
{
size_t i;
int last_good_net = -1;
virDomainNetDefPtr net;
for (i = 0; i < def->nnets; i++) {
net = def->nets[i];
if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) {
if (virNetDevVPortProfileAssociate(net->ifname,
virDomainNetGetActualVirtPortProfile(net),
&net->mac,
virDomainNetGetActualDirectDev(net),
-1,
def->uuid,
VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH,
false) < 0) {
virReportError(VIR_ERR_OPERATION_FAILED,
_("Port profile Associate failed for %s"),
net->ifname);
goto err_exit;
}
last_good_net = i;
VIR_DEBUG("Port profile Associate succeeded for %s", net->ifname);
if (virNetDevMacVLanVPortProfileRegisterCallback(net->ifname, &net->mac,
virDomainNetGetActualDirectDev(net), def->uuid,
virDomainNetGetActualVirtPortProfile(net),
VIR_NETDEV_VPORT_PROFILE_OP_CREATE))
goto err_exit;
}
}
return 0;
err_exit:
for (i = 0; last_good_net != -1 && i <= last_good_net; i++) {
net = def->nets[i];
if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) {
ignore_value(virNetDevVPortProfileDisassociate(net->ifname,
virDomainNetGetActualVirtPortProfile(net),
&net->mac,
virDomainNetGetActualDirectDev(net),
-1,
VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH));
}
}
return -1;
}
static int
qemuMigrationPersist(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMigrationCookiePtr mig,
bool ignoreSaveError)
{
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
virCapsPtr caps = NULL;
virDomainDefPtr vmdef;
virDomainDefPtr oldDef = NULL;
unsigned int oldPersist = vm->persistent;
virObjectEventPtr event;
int ret = -1;
if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
goto cleanup;
vm->persistent = 1;
oldDef = vm->newDef;
vm->newDef = qemuMigrationCookieGetPersistent(mig);
if (!(vmdef = virDomainObjGetPersistentDef(caps, driver->xmlopt, vm)))
goto error;
if (virDomainSaveConfig(cfg->configDir, driver->caps, vmdef) < 0 &&
!ignoreSaveError)
goto error;
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_DEFINED,
oldPersist ?
VIR_DOMAIN_EVENT_DEFINED_UPDATED :
VIR_DOMAIN_EVENT_DEFINED_ADDED);
qemuDomainEventQueue(driver, event);
ret = 0;
cleanup:
virDomainDefFree(oldDef);
virObjectUnref(caps);
virObjectUnref(cfg);
return ret;
error:
virDomainDefFree(vm->newDef);
vm->persistent = oldPersist;
vm->newDef = oldDef;
oldDef = NULL;
goto cleanup;
}
virDomainPtr
qemuMigrationFinish(virQEMUDriverPtr driver,
virConnectPtr dconn,
virDomainObjPtr vm,
const char *cookiein,
int cookieinlen,
char **cookieout,
int *cookieoutlen,
unsigned long flags,
int retcode,
bool v3proto)
{
virDomainPtr dom = NULL;
qemuMigrationCookiePtr mig = NULL;
virErrorPtr orig_err = NULL;
int cookie_flags = 0;
qemuDomainObjPrivatePtr priv = vm->privateData;
virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
unsigned short port;
unsigned long long timeReceived = 0;
virObjectEventPtr event;
int rc;
qemuDomainJobInfoPtr jobInfo = NULL;
bool inPostCopy = false;
bool kill = true;
VIR_DEBUG("driver=%p, dconn=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
"cookieout=%p, cookieoutlen=%p, flags=%lx, retcode=%d",
driver, dconn, vm, NULLSTR(cookiein), cookieinlen,
cookieout, cookieoutlen, flags, retcode);
port = priv->migrationPort;
priv->migrationPort = 0;
if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_IN)) {
qemuMigrationErrorReport(driver, vm->def->name);
goto cleanup;
}
ignore_value(virTimeMillisNow(&timeReceived));
qemuMigrationJobStartPhase(driver, vm,
v3proto ? QEMU_MIGRATION_PHASE_FINISH3
: QEMU_MIGRATION_PHASE_FINISH2);
qemuDomainCleanupRemove(vm, qemuMigrationPrepareCleanup);
VIR_FREE(priv->job.completed);
cookie_flags = QEMU_MIGRATION_COOKIE_NETWORK |
QEMU_MIGRATION_COOKIE_STATS |
QEMU_MIGRATION_COOKIE_NBD;
if (flags & VIR_MIGRATE_PERSIST_DEST)
cookie_flags |= QEMU_MIGRATION_COOKIE_PERSISTENT;
if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein,
cookieinlen, cookie_flags)))
goto endjob;
if (flags & VIR_MIGRATE_OFFLINE) {
if (retcode == 0 &&
qemuMigrationPersist(driver, vm, mig, false) == 0)
dom = virGetDomain(dconn, vm->def->name, vm->def->uuid);
goto endjob;
}
if (retcode != 0) {
qemuDomainJobInfo info;
/* Check for a possible error on the monitor in case Finish was called
* earlier than monitor EOF handler got a chance to process the error
*/
qemuMigrationFetchJobStatus(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_IN,
&info);
goto endjob;
}
if (!virDomainObjIsActive(vm)) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("guest unexpectedly quit"));
qemuMigrationErrorReport(driver, vm->def->name);
goto endjob;
}
if (qemuMigrationVPAssociatePortProfiles(vm->def) < 0)
goto endjob;
if (mig->network && qemuDomainMigrateOPDRelocate(driver, vm, mig) < 0)
VIR_WARN("unable to provide network data for relocation");
if (qemuMigrationStopNBDServer(driver, vm, mig) < 0)
goto endjob;
if (qemuRefreshVirtioChannelState(driver, vm) < 0)
goto endjob;
if ((rc = qemuConnectAgent(driver, vm)) < 0) {
if (rc == -2)
goto endjob;
VIR_WARN("Cannot connect to QEMU guest agent for %s",
vm->def->name);
virResetLastError();
priv->agentError = true;
}
if (flags & VIR_MIGRATE_PERSIST_DEST) {
if (qemuMigrationPersist(driver, vm, mig, !v3proto) < 0) {
/* Hmpf. Migration was successful, but making it persistent
* was not. If we report successful, then when this domain
* shuts down, management tools are in for a surprise. On the
* other hand, if we report failure, then the management tools
* might try to restart the domain on the source side, even
* though the domain is actually running on the destination.
* Pretend success and hope that this is a rare situation and
* management tools are smart.
*
* However, in v3 protocol, the source VM is still available
* to restart during confirm() step, so we kill it off now.
*/
if (v3proto)
goto endjob;
}
}
/* We need to wait for QEMU to process all data sent by the source
* before starting guest CPUs.
*/
if (qemuMigrationWaitForDestCompletion(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_IN,
!!(flags & VIR_MIGRATE_POSTCOPY)) < 0) {
/* There's not much we can do for v2 protocol since the
* original domain on the source host is already gone.
*/
if (v3proto)
goto endjob;
}
if (priv->job.current->stats.status == QEMU_MONITOR_MIGRATION_STATUS_POSTCOPY)
inPostCopy = true;
if (!(flags & VIR_MIGRATE_PAUSED)) {
/* run 'cont' on the destination, which allows migration on qemu
* >= 0.10.6 to work properly. This isn't strictly necessary on
* older qemu's, but it also doesn't hurt anything there
*/
if (qemuProcessStartCPUs(driver, vm, dconn,
inPostCopy ? VIR_DOMAIN_RUNNING_POSTCOPY
: VIR_DOMAIN_RUNNING_MIGRATED,
QEMU_ASYNC_JOB_MIGRATION_IN) < 0) {
if (virGetLastError() == NULL)
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("resume operation failed"));
/* Need to save the current error, in case shutting
* down the process overwrites it
*/
orig_err = virSaveLastError();
/*
* In v3 protocol, the source VM is still available to
* restart during confirm() step, so we kill it off
* now.
* In v2 protocol, the source is dead, so we leave
* target in paused state, in case admin can fix
* things up.
*/
if (v3proto)
goto endjob;
}
if (inPostCopy) {
kill = false;
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_RESUMED,
VIR_DOMAIN_EVENT_RESUMED_POSTCOPY);
qemuDomainEventQueue(driver, event);
}
}
if (mig->jobInfo) {
jobInfo = mig->jobInfo;
mig->jobInfo = NULL;
if (jobInfo->sent && timeReceived) {
jobInfo->timeDelta = timeReceived - jobInfo->sent;
jobInfo->received = timeReceived;
jobInfo->timeDeltaSet = true;
}
qemuDomainJobInfoUpdateTime(jobInfo);
qemuDomainJobInfoUpdateDowntime(jobInfo);
}
if (inPostCopy) {
if (qemuMigrationWaitForDestCompletion(driver, vm,
QEMU_ASYNC_JOB_MIGRATION_IN,
false) < 0) {
goto endjob;
}
if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
virDomainObjSetState(vm,
VIR_DOMAIN_RUNNING,
VIR_DOMAIN_RUNNING_MIGRATED);
}
}
dom = virGetDomain(dconn, vm->def->name, vm->def->uuid);
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_RESUMED,
VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
qemuDomainEventQueue(driver, event);
if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED) {
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_USER);
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_SUSPENDED,
VIR_DOMAIN_EVENT_SUSPENDED_PAUSED);
qemuDomainEventQueue(driver, event);
}
if (virDomainObjIsActive(vm) &&
virDomainSaveStatus(driver->xmlopt, cfg->stateDir, vm, driver->caps) < 0)
VIR_WARN("Failed to save status on vm %s", vm->def->name);
/* Guest is successfully running, so cancel previous auto destroy */
qemuProcessAutoDestroyRemove(driver, vm);
endjob:
if (!dom &&
!(flags & VIR_MIGRATE_OFFLINE) &&
virDomainObjIsActive(vm)) {
if (kill) {
qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED,
QEMU_ASYNC_JOB_MIGRATION_IN,
VIR_QEMU_PROCESS_STOP_MIGRATED);
virDomainAuditStop(vm, "failed");
event = virDomainEventLifecycleNewFromObj(vm,
VIR_DOMAIN_EVENT_STOPPED,
VIR_DOMAIN_EVENT_STOPPED_FAILED);
qemuDomainEventQueue(driver, event);
} else {
qemuMigrationPostcopyFailed(driver, vm);
}
}
if (dom) {
priv->job.completed = jobInfo;
jobInfo = NULL;
if (qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen,
QEMU_MIGRATION_COOKIE_STATS) < 0)
VIR_WARN("Unable to encode migration cookie");
/* Remove completed stats for post-copy, everything but timing fields
* is obsolete anyway.
*/
if (inPostCopy)
VIR_FREE(priv->job.completed);
}
qemuMigrationJobFinish(driver, vm);
if (!virDomainObjIsActive(vm))
qemuDomainRemoveInactive(driver, vm);
cleanup:
VIR_FREE(jobInfo);
virPortAllocatorRelease(driver->migrationPorts, port);
if (priv->mon)
qemuMonitorSetDomainLog(priv->mon, NULL, NULL, NULL);
VIR_FREE(priv->origname);
virDomainObjEndAPI(&vm);
if (mig) {
virDomainDefFree(qemuMigrationCookieGetPersistent(mig));
qemuMigrationCookieFree(mig);
}
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
virObjectUnref(cfg);
/* Set a special error if Finish is expected to return NULL as a result of
* successful call with retcode != 0
*/
if (retcode != 0 && !dom && !virGetLastError())
virReportError(VIR_ERR_MIGRATE_FINISH_OK, NULL);
return dom;
}
/* Helper function called while vm is active. */
int
qemuMigrationToFile(virQEMUDriverPtr driver, virDomainObjPtr vm,
int fd,
const char *compressor,
qemuDomainAsyncJob asyncJob)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int rc;
int ret = -1;
virCommandPtr cmd = NULL;
int pipeFD[2] = { -1, -1 };
unsigned long saveMigBandwidth = priv->migMaxBandwidth;
char *errbuf = NULL;
virErrorPtr orig_err = NULL;
/* Increase migration bandwidth to unlimited since target is a file.
* Failure to change migration speed is not fatal. */
if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) == 0) {
qemuMonitorSetMigrationSpeed(priv->mon,
QEMU_DOMAIN_MIG_BANDWIDTH_MAX);
priv->migMaxBandwidth = QEMU_DOMAIN_MIG_BANDWIDTH_MAX;
if (qemuDomainObjExitMonitor(driver, vm) < 0)
return -1;
}
if (!virDomainObjIsActive(vm)) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("guest unexpectedly quit"));
/* nothing to tear down */
return -1;
}
if (compressor && pipe(pipeFD) < 0) {
virReportSystemError(errno, "%s",
_("Failed to create pipe for migration"));
return -1;
}
/* All right! We can use fd migration, which means that qemu
* doesn't have to open() the file, so while we still have to
* grant SELinux access, we can do it on fd and avoid cleanup
* later, as well as skip futzing with cgroup. */
if (virSecurityManagerSetImageFDLabel(driver->securityManager, vm->def,
compressor ? pipeFD[1] : fd) < 0)
goto cleanup;
if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
goto cleanup;
if (!compressor) {
rc = qemuMonitorMigrateToFd(priv->mon,
QEMU_MONITOR_MIGRATE_BACKGROUND,
fd);
} else {
const char *prog = compressor;
const char *args[] = {
prog,
"-c",
NULL
};
cmd = virCommandNewArgs(args);
virCommandSetInputFD(cmd, pipeFD[0]);
virCommandSetOutputFD(cmd, &fd);
virCommandSetErrorBuffer(cmd, &errbuf);
virCommandDoAsyncIO(cmd);
if (virSetCloseExec(pipeFD[1]) < 0) {
virReportSystemError(errno, "%s",
_("Unable to set cloexec flag"));
ignore_value(qemuDomainObjExitMonitor(driver, vm));
goto cleanup;
}
if (virCommandRunAsync(cmd, NULL) < 0) {
ignore_value(qemuDomainObjExitMonitor(driver, vm));
goto cleanup;
}
rc = qemuMonitorMigrateToFd(priv->mon,
QEMU_MONITOR_MIGRATE_BACKGROUND,
pipeFD[1]);
if (VIR_CLOSE(pipeFD[0]) < 0 ||
VIR_CLOSE(pipeFD[1]) < 0)
VIR_WARN("failed to close intermediate pipe");
}
if (qemuDomainObjExitMonitor(driver, vm) < 0)
goto cleanup;
if (rc < 0)
goto cleanup;
rc = qemuMigrationWaitForCompletion(driver, vm, asyncJob,
NULL, false, false);
if (rc < 0) {
if (rc == -2) {
orig_err = virSaveLastError();
virCommandAbort(cmd);
if (virDomainObjIsActive(vm) &&
qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) == 0) {
qemuMonitorMigrateCancel(priv->mon);
ignore_value(qemuDomainObjExitMonitor(driver, vm));
}
}
goto cleanup;
}
if (cmd && virCommandWait(cmd, NULL) < 0)
goto cleanup;
qemuDomainEventEmitJobCompleted(driver, vm);
ret = 0;
cleanup:
if (ret < 0 && !orig_err)
orig_err = virSaveLastError();
/* Restore max migration bandwidth */
if (virDomainObjIsActive(vm) &&
qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) == 0) {
qemuMonitorSetMigrationSpeed(priv->mon, saveMigBandwidth);
priv->migMaxBandwidth = saveMigBandwidth;
ignore_value(qemuDomainObjExitMonitor(driver, vm));
}
VIR_FORCE_CLOSE(pipeFD[0]);
VIR_FORCE_CLOSE(pipeFD[1]);
if (cmd) {
VIR_DEBUG("Compression binary stderr: %s", NULLSTR(errbuf));
VIR_FREE(errbuf);
virCommandFree(cmd);
}
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
return ret;
}
int
qemuMigrationCancel(virQEMUDriverPtr driver,
virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
virHashTablePtr blockJobs = NULL;
bool storage = false;
size_t i;
int ret = -1;
VIR_DEBUG("Canceling unfinished outgoing migration of domain %s",
vm->def->name);
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
if (QEMU_DOMAIN_DISK_PRIVATE(disk)->migrating) {
qemuBlockJobSyncBegin(disk);
storage = true;
}
}
qemuDomainObjEnterMonitor(driver, vm);
ignore_value(qemuMonitorMigrateCancel(priv->mon));
if (storage)
blockJobs = qemuMonitorGetAllBlockJobInfo(priv->mon);
if (qemuDomainObjExitMonitor(driver, vm) < 0 || (storage && !blockJobs))
goto endsyncjob;
if (!storage) {
ret = 0;
goto cleanup;
}
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
if (!diskPriv->migrating)
continue;
if (virHashLookup(blockJobs, disk->info.alias)) {
VIR_DEBUG("Drive mirror on disk %s is still running", disk->dst);
} else {
VIR_DEBUG("Drive mirror on disk %s is gone", disk->dst);
qemuBlockJobSyncEnd(driver, vm, disk);
diskPriv->migrating = false;
}
}
if (qemuMigrationCancelDriveMirror(driver, vm, false,
QEMU_ASYNC_JOB_NONE, NULL) < 0)
goto endsyncjob;
ret = 0;
cleanup:
virHashFree(blockJobs);
return ret;
endsyncjob:
if (storage) {
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk = vm->def->disks[i];
qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
if (diskPriv->migrating) {
qemuBlockJobSyncEnd(driver, vm, disk);
diskPriv->migrating = false;
}
}
}
goto cleanup;
}
int
qemuMigrationJobStart(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuDomainAsyncJob job)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (qemuDomainObjBeginAsyncJob(driver, vm, job) < 0)
return -1;
if (job == QEMU_ASYNC_JOB_MIGRATION_IN) {
qemuDomainObjSetAsyncJobMask(vm, QEMU_JOB_NONE);
} else {
qemuDomainObjSetAsyncJobMask(vm, (QEMU_JOB_DEFAULT_MASK |
JOB_MASK(QEMU_JOB_SUSPEND) |
JOB_MASK(QEMU_JOB_MIGRATION_OP)));
}
priv->job.current->type = VIR_DOMAIN_JOB_UNBOUNDED;
return 0;
}
void
qemuMigrationJobSetPhase(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMigrationJobPhase phase)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (phase < priv->job.phase) {
VIR_ERROR(_("migration protocol going backwards %s => %s"),
qemuMigrationJobPhaseTypeToString(priv->job.phase),
qemuMigrationJobPhaseTypeToString(phase));
return;
}
qemuDomainObjSetJobPhase(driver, vm, phase);
}
void
qemuMigrationJobStartPhase(virQEMUDriverPtr driver,
virDomainObjPtr vm,
qemuMigrationJobPhase phase)
{
qemuMigrationJobSetPhase(driver, vm, phase);
}
void
qemuMigrationJobContinue(virDomainObjPtr vm)
{
qemuDomainObjReleaseAsyncJob(vm);
}
bool
qemuMigrationJobIsActive(virDomainObjPtr vm,
qemuDomainAsyncJob job)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (priv->job.asyncJob != job) {
const char *msg;
if (job == QEMU_ASYNC_JOB_MIGRATION_IN)
msg = _("domain '%s' is not processing incoming migration");
else
msg = _("domain '%s' is not being migrated");
virReportError(VIR_ERR_OPERATION_INVALID, msg, vm->def->name);
return false;
}
return true;
}
void
qemuMigrationJobFinish(virQEMUDriverPtr driver, virDomainObjPtr vm)
{
qemuDomainObjEndAsyncJob(driver, vm);
}
static void
qemuMigrationErrorFree(void *data,
const void *name ATTRIBUTE_UNUSED)
{
virErrorPtr err = data;
virFreeError(err);
}
int
qemuMigrationErrorInit(virQEMUDriverPtr driver)
{
driver->migrationErrors = virHashAtomicNew(64, qemuMigrationErrorFree);
if (driver->migrationErrors)
return 0;
else
return -1;
}
/**
* This function consumes @err; the caller should consider the @err pointer
* invalid after calling this function.
*/
void
qemuMigrationErrorSave(virQEMUDriverPtr driver,
const char *name,
virErrorPtr err)
{
if (!err)
return;
VIR_DEBUG("Saving incoming migration error for domain %s: %s",
name, err->message);
if (virHashAtomicUpdate(driver->migrationErrors, name, err) < 0) {
VIR_WARN("Failed to save migration error for domain '%s'", name);
virFreeError(err);
}
}
void
qemuMigrationErrorReport(virQEMUDriverPtr driver,
const char *name)
{
virErrorPtr err;
if (!(err = virHashAtomicSteal(driver->migrationErrors, name)))
return;
VIR_DEBUG("Restoring saved incoming migration error for domain %s: %s",
name, err->message);
virSetError(err);
virFreeError(err);
}