Skip to content

T
TDengine

taosdata / TDengine
大约 1 年 前同步成功

通知 1185
Star 22015
Fork 4786
T
TDengine
提交 70625808 编写于 作者: T Tao Liu
浏览文件
操作

Merge branch 'develop' of https://github.com/taosdata/TDengine into develop

上级 22d60979 e6ba87bd
隐藏空白更改
内联 并排
Showing with 2197 addition and 810 deletion
.travis.yml
浏览文件 @ 70625808
......@@ -13,7 +13,7 @@ branches:
matrix:
- os: linux
dist: bionic
dist: focal
language: c
git:
......@@ -28,8 +28,6 @@ matrix:
- build-essential
- cmake
- net-tools
- python-pip
- python-setuptools
- python3-pip
- python3-setuptools
- valgrind
......@@ -54,13 +52,19 @@ matrix:
cd ${TRAVIS_BUILD_DIR}/debug
make install > /dev/null || travis_terminate $?
pip install numpy
pip install --user ${TRAVIS_BUILD_DIR}/src/connector/python/linux/python2/
pip3 install numpy
pip3 install --user ${TRAVIS_BUILD_DIR}/src/connector/python/linux/python3/
cd ${TRAVIS_BUILD_DIR}/tests
./test-all.sh smoke || travis_terminate $?
sleep 1
cd ${TRAVIS_BUILD_DIR}/tests/pytest
pkill -TERM -x taosd
fuser -k -n tcp 6030
sleep 1
./crash_gen.sh -a -p -t 4 -s 25|| travis_terminate $?
sleep 1
cd ${TRAVIS_BUILD_DIR}/tests/pytest
./valgrind-test.sh 2>&1 > mem-error-out.log
......@@ -160,7 +164,7 @@ matrix:
script:
- cmake .. > /dev/null
- make > /dev/null
- make
- os: linux
dist: bionic
......
src/client/src/tscLocalMerge.c
浏览文件 @ 70625808
......@@ -274,6 +274,10 @@ void tscCreateLocalReducer(tExtMemBuffer **pMemBuffer, int32_t numOfBuffer, tOrd
pReducer->numOfBuffer = idx;
SCompareParam *param = malloc(sizeof(SCompareParam));
if (param == NULL) {
tfree(pReducer);
return;
}
param->pLocalData = pReducer->pLocalDataSrc;
param->pDesc = pReducer->pDesc;
param->num = pReducer->pLocalDataSrc[0]->pMemBuffer->numOfElemsPerPage;
......@@ -284,6 +288,7 @@ void tscCreateLocalReducer(tExtMemBuffer **pMemBuffer, int32_t numOfBuffer, tOrd
pRes->code = tLoserTreeCreate(&pReducer->pLoserTree, pReducer->numOfBuffer, param, treeComparator);
if (pReducer->pLoserTree == NULL || pRes->code != 0) {
tfree(param);
tfree(pReducer);
return;
}
......@@ -332,6 +337,8 @@ void tscCreateLocalReducer(tExtMemBuffer **pMemBuffer, int32_t numOfBuffer, tOrd
tfree(pReducer->pResultBuf);
tfree(pReducer->pFinalRes);
tfree(pReducer->prevRowOfInput);
tfree(pReducer->pLoserTree);
tfree(param);
tfree(pReducer);
pRes->code = TSDB_CODE_TSC_OUT_OF_MEMORY;
return;
......
src/client/src/tscStream.c
浏览文件 @ 70625808
......@@ -255,6 +255,9 @@ static void tscProcessStreamRetrieveResult(void *param, TAOS_RES *res, int numOf
// release the metric/meter meta information reference, so data in cache can be updated
taosCacheRelease(tscCacheHandle, (void**)&(pTableMetaInfo->pTableMeta), false);
tscFreeSqlResult(pSql);
tfree(pSql->pSubs);
pSql->numOfSubs = 0;
tfree(pTableMetaInfo->vgroupList);
tscSetNextLaunchTimer(pStream, pSql);
}
......
src/client/src/tscSubquery.c
浏览文件 @ 70625808
......@@ -1447,9 +1447,7 @@ int32_t tscHandleMasterSTableQuery(SSqlObj *pSql) {
static void tscFreeSubSqlObj(SRetrieveSupport *trsupport, SSqlObj *pSql) {
tscDebug("%p start to free subquery result", pSql);
if (pSql->res.code == TSDB_CODE_SUCCESS) {
taos_free_result(pSql);
}
taos_free_result(pSql);
tfree(trsupport->localBuffer);
......@@ -1780,6 +1778,7 @@ static SSqlObj *tscCreateSqlObjForSubquery(SSqlObj *pSql, SRetrieveSupport *trsu
pSql->pSubs[trsupport->subqueryIndex] = pNew;
}
printf("------------alloc:%p\n", pNew);
return pNew;
}
......
src/client/src/tscUtil.c
浏览文件 @ 70625808
......@@ -1822,7 +1822,6 @@ SSqlObj* createSubqueryObj(SSqlObj* pSql, int16_t tableIndex, void (*fp)(), void
STableMeta* pPrevTableMeta = taosCacheTransfer(tscCacheHandle, (void**)&pPrevInfo->pTableMeta);
SVgroupsInfo* pVgroupsInfo = pPrevInfo->vgroupList;
pPrevInfo->vgroupList = NULL;
pFinalInfo = tscAddTableMetaInfo(pNewQueryInfo, name, pPrevTableMeta, pVgroupsInfo, pTableMetaInfo->tagColList);
}
......
src/common/src/tglobal.c
浏览文件 @ 70625808
......@@ -129,7 +129,7 @@ int32_t tsMnodeEqualVnodeNum = 4;
int32_t tsEnableHttpModule = 1;
int32_t tsRestRowLimit = 10240;
uint16_t tsHttpPort = 6020; // only tcp, range tcp[6020]
int32_t tsHttpCacheSessions = 100;
int32_t tsHttpCacheSessions = 1000;
int32_t tsHttpSessionExpire = 36000;
int32_t tsHttpMaxThreads = 2;
int32_t tsHttpEnableCompress = 0;
......
src/connector/jdbc/buildTDengine.sh 已删除 100755 → 0
浏览文件 @ 22d60979
#!/bin/bash
ulimit -c unlimited
function buildTDengine {
cd /root/TDengine
git remote update
REMOTE_COMMIT=`git rev-parse --short remotes/origin/develop`
LOCAL_COMMIT=`git rev-parse --short @`
echo " LOCAL: $LOCAL_COMMIT"
echo "REMOTE: $REMOTE_COMMIT"
if [ "$LOCAL_COMMIT" == "$REMOTE_COMMIT" ]; then
echo "repo up-to-date"
else
echo "repo need to pull"
git pull
LOCAL_COMMIT=`git rev-parse --short @`
cd /root/TDengine/debug
rm -rf /root/TDengine/debug/*
cmake ..
make > /dev/null
make install
fi
}
function restartTaosd {
systemctl stop taosd
pkill -KILL -x taosd
sleep 10
logDir=`grep 'logDir' /etc/taos/taos.cfg|awk 'END{print $2}'`
dataDir=`grep 'dataDir' /etc/taos/taos.cfg|awk '{print $2}'`
rm -rf $logDir/*
rm -rf $dataDir/*
taosd 2>&1 > /dev/null &
sleep 10
}
buildTDengine
restartTaosd
src/connector/jdbc/src/main/java/com/taosdata/jdbc/utils/TDNode.java 0 → 100644
浏览文件 @ 70625808
package com.taosdata.jdbc.utils;
import java.io.File;
import java.util.*;
import java.util.concurrent.TimeUnit;
public class TDNode {
private int index;
private int running;
private int deployed;
private boolean testCluster;
private String path;
private String cfgDir;
private String dataDir;
private String logDir;
private String cfgPath;
public TDNode(int index) {
this.index = index;
running = 0;
deployed = 0;
testCluster = false;
}
public void setPath(String path) {
this.path = path;
}
public void setTestCluster(boolean testCluster) {
this.testCluster = testCluster;
}
public void searchTaosd(File dir, ArrayList<String> taosdPath) {
File[] fileList = dir.listFiles();
if(fileList == null || fileList.length == 0) {
return;
}
for(File file : fileList) {
if(file.isFile()) {
if(file.getName().equals("taosd")) {
taosdPath.add(file.getAbsolutePath());
}
} else {
searchTaosd(file, taosdPath);
}
}
}
public void start() {
String selfPath = System.getProperty("user.dir");
String binPath = "";
String projDir = selfPath + "/../../../";
try {
ArrayList<String> taosdPath = new ArrayList<>();
File dir = new File(projDir);
String realProjDir = dir.getCanonicalPath();
dir = new File(realProjDir);
System.out.println("project Dir: " + projDir);
searchTaosd(dir, taosdPath);
if(taosdPath.size() == 0) {
System.out.println("The project path doens't exist");
return;
} else {
for(String p : taosdPath) {
if(!p.contains("packaging")) {
binPath = p;
}
}
}
} catch (Exception e) {
e.printStackTrace();
}
if(binPath.isEmpty()) {
System.out.println("taosd not found");
return;
} else {
System.out.println("taosd found in " + binPath);
}
if(this.deployed == 0) {
System.out.println("dnode" + index + "is not deployed");
return;
}
String cmd = "nohup " + binPath + " -c " + cfgDir + " > /dev/null 2>&1 & ";
System.out.println("start taosd cmd: " + cmd);
try{
Runtime.getRuntime().exec(cmd);
TimeUnit.SECONDS.sleep(5);
} catch (Exception e) {
e.printStackTrace();
}
this.running = 1;
}
public void stop() {
String toBeKilled = "taosd";
if (this.running != 0) {
String killCmd = "pkill -kill -x " + toBeKilled;
String[] killCmds = {"sh", "-c", killCmd};
try {
Runtime.getRuntime().exec(killCmds).waitFor();
for(int port = 6030; port < 6041; port ++) {
String fuserCmd = "fuser -k -n tcp " + port;
Runtime.getRuntime().exec(fuserCmd).waitFor();
}
} catch (Exception e) {
e.printStackTrace();
}
this.running = 0;
System.out.println("dnode:" + this.index + " is stopped by pkill");
}
}
public void startIP() {
try{
String cmd = "sudo ifconfig lo:" + index + "192.168.0." + index + " up";
Runtime.getRuntime().exec(cmd).waitFor();
} catch (Exception e) {
e.printStackTrace();
}
}
public void stopIP() {
try{
String cmd = "sudo ifconfig lo:" + index + "192.168.0." + index + " down";
Runtime.getRuntime().exec(cmd).waitFor();
} catch (Exception e) {
e.printStackTrace();
}
}
public void setCfgConfig(String option, String value) {
try{
String cmd = "echo " + option + " " + value + " >> " + this.cfgPath;
String[] cmdLine = {"sh", "-c", cmd};
Process ps = Runtime.getRuntime().exec(cmdLine);
ps.waitFor();
} catch (Exception e) {
e.printStackTrace();
}
}
public String getDnodeRootDir() {
String dnodeRootDir = this.path + "/sim/psim/dnode" + this.index;
return dnodeRootDir;
}
public String getDnodesRootDir() {
String dnodesRootDir = this.path + "/sim/psim" + this.index;
return dnodesRootDir;
}
public void deploy() {
this.logDir = this.path + "/sim/dnode" + this.index + "/log";
this.dataDir = this.path + "/sim/dnode" + this.index + "/data";
this.cfgDir = this.path + "/sim/dnode" + this.index + "/cfg";
this.cfgPath = this.path + "/sim/dnode" + this.index + "/cfg/taos.cfg";
try {
String cmd = "rm -rf " + this.logDir;
Runtime.getRuntime().exec(cmd).waitFor();
cmd = "rm -rf " + this.cfgDir;
Runtime.getRuntime().exec(cmd).waitFor();
cmd = "rm -rf " + this.dataDir;
Runtime.getRuntime().exec(cmd).waitFor();
cmd = "mkdir -p " + this.logDir;
Runtime.getRuntime().exec(cmd).waitFor();
cmd = "mkdir -p " + this.cfgDir;
Runtime.getRuntime().exec(cmd).waitFor();
cmd = "mkdir -p " + this.dataDir;
Runtime.getRuntime().exec(cmd).waitFor();
cmd = "touch " + this.cfgPath;
Runtime.getRuntime().exec(cmd).waitFor();
} catch (Exception e) {
e.printStackTrace();
}
if(this.testCluster) {
startIP();
setCfgConfig("masterIp", "192.168.0.1");
setCfgConfig("secondIp", "192.168.0.2");
setCfgConfig("publicIp", "192.168.0." + this.index);
setCfgConfig("internalIp", "192.168.0." + this.index);
setCfgConfig("privateIp", "192.168.0." + this.index);
}
setCfgConfig("dataDir", this.dataDir);
setCfgConfig("logDir", this.logDir);
setCfgConfig("numOfLogLines", "1000000/00");
setCfgConfig("mnodeEqualVnodeNum", "0");
setCfgConfig("walLevel", "1");
setCfgConfig("statusInterval", "1");
setCfgConfig("numOfTotalVnodes", "64");
setCfgConfig("numOfMnodes", "3");
setCfgConfig("numOfThreadsPerCore", "2.0");
setCfgConfig("monitor", "0");
setCfgConfig("maxVnodeConnections", "30000");
setCfgConfig("maxMgmtConnections", "30000");
setCfgConfig("maxMeterConnections", "30000");
setCfgConfig("maxShellConns", "30000");
setCfgConfig("locale", "en_US.UTF-8");
setCfgConfig("charset", "UTF-8");
setCfgConfig("asyncLog", "0");
setCfgConfig("anyIp", "0");
setCfgConfig("dDebugFlag", "135");
setCfgConfig("mDebugFlag", "135");
setCfgConfig("sdbDebugFlag", "135");
setCfgConfig("rpcDebugFlag", "135");
setCfgConfig("tmrDebugFlag", "131");
setCfgConfig("cDebugFlag", "135");
setCfgConfig("httpDebugFlag", "135");
setCfgConfig("monitorDebugFlag", "135");
setCfgConfig("udebugFlag", "135");
setCfgConfig("jnidebugFlag", "135");
setCfgConfig("qdebugFlag", "135");
this.deployed = 1;
}
}
\ No newline at end of file
src/connector/jdbc/src/main/java/com/taosdata/jdbc/utils/TDNodes.java 0 → 100644
浏览文件 @ 70625808
package com.taosdata.jdbc.utils;
import java.io.File;
import java.util.*;
public class TDNodes {
private ArrayList<TDNode> tdNodes;
private boolean testCluster;
public TDNodes () {
tdNodes = new ArrayList<>();
for(int i = 1; i < 11; i ++) {
tdNodes.add(new TDNode(i));
}
}
public void setPath(String path) {
try {
String killCmd = "pkill -kill -x taosd";
String[] killCmds = {"sh", "-c", killCmd};
Runtime.getRuntime().exec(killCmds).waitFor();
String binPath = System.getProperty("user.dir");
binPath += "/../../../debug";
System.out.println("binPath: " + binPath);
File file = new File(path);
binPath = file.getCanonicalPath();
System.out.println("binPath real path: " + binPath);
if(path.isEmpty()){
file = new File(path + "/../../");
path = file.getCanonicalPath();
}
for(int i = 0; i < tdNodes.size(); i++) {
tdNodes.get(i).setPath(path);
}
} catch (Exception e) {
e.printStackTrace();
}
}
public void setTestCluster(boolean testCluster) {
this.testCluster = testCluster;
}
public void check(int index) {
if(index < 1 || index > 10) {
System.out.println("index: " + index + " should on a scale of [1, 10]");
return;
}
}
public void deploy(int index) {
try {
File file = new File(System.getProperty("user.dir") + "/../../../");
String projectRealPath = file.getCanonicalPath();
check(index);
tdNodes.get(index - 1).setTestCluster(this.testCluster);
tdNodes.get(index - 1).setPath(projectRealPath);
tdNodes.get(index - 1).deploy();
} catch (Exception e) {
e.printStackTrace();
System.out.println("deploy Test Exception");
}
}
public void cfg(int index, String option, String value) {
check(index);
tdNodes.get(index - 1).setCfgConfig(option, value);
}
public void start(int index) {
check(index);
tdNodes.get(index - 1).start();
}
public void stop(int index) {
check(index);
tdNodes.get(index - 1).stop();
}
public void startIP(int index) {
check(index);
tdNodes.get(index - 1).startIP();
}
public void stopIP(int index) {
check(index);
tdNodes.get(index - 1).stopIP();
}
}
\ No newline at end of file
src/connector/jdbc/src/test/java/com/taosdata/jdbc/BaseTest.java
浏览文件 @ 70625808
package com.taosdata.jdbc;
import java.io.BufferedReader;
import java.io.InputStreamReader;
import java.io.File;
import com.taosdata.jdbc.utils.TDNodes;
import org.junit.AfterClass;
import org.junit.BeforeClass;
public class BaseTest {
private static boolean testCluster = false;
private static String deployPath = System.getProperty("user.dir");
private static TDNodes tdNodes = new TDNodes();
@BeforeClass
public static void setupEnv() {
try{
String path = System.getProperty("user.dir");
String bashPath = path + "/buildTDengine.sh";
File file = new File(deployPath + "/../../../");
String rootPath = file.getCanonicalPath();
tdNodes.setPath(rootPath);
tdNodes.setTestCluster(testCluster);
Process ps = Runtime.getRuntime().exec(bashPath);
ps.waitFor();
tdNodes.deploy(1);
tdNodes.start(1);
BufferedReader br = new BufferedReader(new InputStreamReader(ps.getInputStream()));
while(br.readLine() != null) {
System.out.println(br.readLine());
}
} catch (Exception e) {
e.printStackTrace();
System.out.println("Base Test Exception");
}
}
@AfterClass
public static void cleanUpEnv() {
tdNodes.stop(1);
}
}
\ No newline at end of file
src/dnode/inc/dnodeMgmt.h
浏览文件 @ 70625808
......@@ -22,6 +22,8 @@ extern "C" {
int32_t dnodeInitMgmt();
void dnodeCleanupMgmt();
int32_t dnodeInitMgmtTimer();
void dnodeCleanupMgmtTimer();
void dnodeDispatchToMgmtQueue(SRpcMsg *rpcMsg);
void* dnodeGetVnode(int32_t vgId);
......
src/dnode/src/dnodeMain.c
浏览文件 @ 70625808
......@@ -57,6 +57,7 @@ static const SDnodeComponent tsDnodeComponents[] = {
{"server", dnodeInitServer, dnodeCleanupServer},
{"mgmt", dnodeInitMgmt, dnodeCleanupMgmt},
{"modules", dnodeInitModules, dnodeCleanupModules},
{"mgmt-tmr",dnodeInitMgmtTimer, dnodeCleanupMgmtTimer},
{"shell", dnodeInitShell, dnodeCleanupShell}
};
......
src/dnode/src/dnodeMgmt.c
浏览文件 @ 70625808
......@@ -147,6 +147,12 @@ int32_t dnodeInitMgmt() {
return -1;
}
dInfo("dnode mgmt is initialized");
return TSDB_CODE_SUCCESS;
}
int32_t dnodeInitMgmtTimer() {
tsDnodeTmr = taosTmrInit(100, 200, 60000, "DND-DM");
if (tsDnodeTmr == NULL) {
dError("failed to init dnode timer");
......@@ -155,13 +161,11 @@ int32_t dnodeInitMgmt() {
}
taosTmrReset(dnodeSendStatusMsg, 500, NULL, tsDnodeTmr, &tsStatusTimer);
dInfo("dnode mgmt is initialized");
dInfo("dnode mgmt timer is initialized");
return TSDB_CODE_SUCCESS;
}
void dnodeCleanupMgmt() {
void dnodeCleanupMgmtTimer() {
if (tsStatusTimer != NULL) {
taosTmrStopA(&tsStatusTimer);
tsStatusTimer = NULL;
......@@ -171,7 +175,10 @@ void dnodeCleanupMgmt() {
taosTmrCleanUp(tsDnodeTmr);
tsDnodeTmr = NULL;
}
}
void dnodeCleanupMgmt() {
dnodeCleanupMgmtTimer();
dnodeCloseVnodes();
if (tsMgmtQset) taosQsetThreadResume(tsMgmtQset);
......
src/dnode/src/dnodeSystem.c
浏览文件 @ 70625808
......@@ -20,6 +20,7 @@
#include "tglobal.h"
#include "dnodeInt.h"
#include "dnodeMain.h"
#include "tfile.h"
static void signal_handler(int32_t signum, siginfo_t *sigInfo, void *context);
static sem_t exitSem;
......@@ -67,6 +68,18 @@ int32_t main(int32_t argc, char *argv[]) {
taosSetAllocMode(TAOS_ALLOC_MODE_DETECT_LEAK, NULL, true);
}
}
#endif
#ifdef TAOS_RANDOM_FILE_FAIL
else if (strcmp(argv[i], "--random-file-fail-factor") == 0) {
if ( (i+1) < argc ) {
int factor = atoi(argv[i+1]);
printf("The factor of random failure is %d\n", factor);
taosSetRandomFileFailFactor(factor);
} else {
printf("Please specify a number for random failure factor!");
exit(EXIT_FAILURE);
}
}
#endif
}
......
src/mnode/src/mnodeDb.c
浏览文件 @ 70625808
......@@ -179,9 +179,14 @@ void mnodeDecDbRef(SDbObj *pDb) {
SDbObj *mnodeGetDbByTableId(char *tableId) {
char db[TSDB_TABLE_ID_LEN], *pos;
// tableId format should be : acct.db.table
pos = strstr(tableId, TS_PATH_DELIMITER);
assert(NULL != pos);
pos = strstr(pos + 1, TS_PATH_DELIMITER);
assert(NULL != pos);
memset(db, 0, sizeof(db));
strncpy(db, tableId, pos - tableId);
......
src/mnode/src/mnodeMain.c
浏览文件 @ 70625808
......@@ -41,7 +41,7 @@ typedef struct {
void (*cleanup)();
} SMnodeComponent;
void *tsMnodeTmr;
void *tsMnodeTmr = NULL;
static bool tsMgmtIsRunning = false;
static const SMnodeComponent tsMnodeComponents[] = {
......
src/mnode/src/mnodeVgroup.c
浏览文件 @ 70625808
......@@ -372,7 +372,6 @@ static int32_t mnodeCreateVgroupCb(SMnodeMsg *pMsg, int32_t code) {
pVgroup->vnodeGid[i].dnodeId);
}
mnodeIncVgroupRef(pVgroup);
pMsg->expected = pVgroup->numOfVnodes;
mnodeSendCreateVgroupMsg(pVgroup, pMsg);
......@@ -393,6 +392,9 @@ int32_t mnodeCreateVgroup(SMnodeMsg *pMsg, SDbObj *pDb) {
return TSDB_CODE_MND_NO_ENOUGH_DNODES;
}
pMsg->pVgroup = pVgroup;
mnodeIncVgroupRef(pVgroup);
SSdbOper oper = {
.type = SDB_OPER_GLOBAL,
.table = tsVgroupSdb,
......@@ -402,8 +404,6 @@ int32_t mnodeCreateVgroup(SMnodeMsg *pMsg, SDbObj *pDb) {
.cb = mnodeCreateVgroupCb
};
pMsg->pVgroup = pVgroup;
int32_t code = sdbInsertRow(&oper);
if (code != TSDB_CODE_SUCCESS) {
pMsg->pVgroup = NULL;
......@@ -814,19 +814,20 @@ static int32_t mnodeProcessVnodeCfgMsg(SMnodeMsg *pMsg) {
mDebug("dnode:%s, vgId:%d, invalid dnode", taosIpStr(pCfg->dnodeId), pCfg->vgId);
return TSDB_CODE_MND_VGROUP_NOT_EXIST;
}
mnodeDecDnodeRef(pDnode);
SVgObj *pVgroup = mnodeGetVgroup(pCfg->vgId);
if (pVgroup == NULL) {
mDebug("dnode:%s, vgId:%d, no vgroup info", taosIpStr(pCfg->dnodeId), pCfg->vgId);
mnodeDecDnodeRef(pDnode);
return TSDB_CODE_MND_VGROUP_NOT_EXIST;
}
mnodeDecVgroupRef(pVgroup);
mDebug("vgId:%d, send create vnode msg to dnode %s for vnode cfg msg", pVgroup->vgId, pDnode->dnodeEp);
SRpcIpSet ipSet = mnodeGetIpSetFromIp(pDnode->dnodeEp);
mnodeSendCreateVnodeMsg(pVgroup, &ipSet, NULL);
mnodeDecDnodeRef(pDnode);
mnodeDecVgroupRef(pVgroup);
return TSDB_CODE_SUCCESS;
}
......
src/plugins/http/inc/httpInt.h
浏览文件 @ 70625808
......@@ -206,7 +206,7 @@ typedef struct HttpThread {
pthread_mutex_t threadMutex;
bool stop;
int pollFd;
int numOfFds;
int numOfContexts;
int threadId;
char label[HTTP_LABEL_SIZE];
bool (*processData)(HttpContext *pContext);
......
src/plugins/http/src/httpContext.c
浏览文件 @ 70625808
......@@ -44,7 +44,7 @@ static void httpDestroyContext(void *data) {
HttpThread *pThread = pContext->pThread;
httpRemoveContextFromEpoll(pContext);
httpReleaseSession(pContext);
atomic_sub_fetch_32(&pThread->numOfFds, 1);
atomic_sub_fetch_32(&pThread->numOfContexts, 1);
pContext->pThread = 0;
pContext->state = HTTP_CONTEXT_STATE_CLOSED;
......@@ -171,38 +171,39 @@ bool httpInitContext(HttpContext *pContext) {
void httpCloseContextByApp(HttpContext *pContext) {
pContext->parsed = false;
bool keepAlive = true;
if (pContext->httpVersion == HTTP_VERSION_10 && pContext->httpKeepAlive != HTTP_KEEPALIVE_ENABLE) {
keepAlive = false;
} else if (pContext->httpVersion != HTTP_VERSION_10 && pContext->httpKeepAlive == HTTP_KEEPALIVE_DISABLE) {
keepAlive = false;
} else {}
} else {
}
if (keepAlive) {
if (httpAlterContextState(pContext, HTTP_CONTEXT_STATE_HANDLING, HTTP_CONTEXT_STATE_READY)) {
httpDebug("context:%p, fd:%d, ip:%s, last state:handling, keepAlive:true, reuse connect",
pContext, pContext->fd, pContext->ipstr);
httpDebug("context:%p, fd:%d, ip:%s, last state:handling, keepAlive:true, reuse context", pContext, pContext->fd,
pContext->ipstr);
} else if (httpAlterContextState(pContext, HTTP_CONTEXT_STATE_DROPPING, HTTP_CONTEXT_STATE_CLOSED)) {
httpRemoveContextFromEpoll(pContext);
httpDebug("context:%p, fd:%d, ip:%s, last state:dropping, keepAlive:true, close connect",
pContext, pContext->fd, pContext->ipstr);
httpDebug("context:%p, fd:%d, ip:%s, last state:dropping, keepAlive:true, close connect", pContext, pContext->fd,
pContext->ipstr);
} else if (httpAlterContextState(pContext, HTTP_CONTEXT_STATE_READY, HTTP_CONTEXT_STATE_READY)) {
httpDebug("context:%p, fd:%d, ip:%s, last state:ready, keepAlive:true, reuse connect",
pContext, pContext->fd, pContext->ipstr);
httpDebug("context:%p, fd:%d, ip:%s, last state:ready, keepAlive:true, reuse context", pContext, pContext->fd,
pContext->ipstr);
} else if (httpAlterContextState(pContext, HTTP_CONTEXT_STATE_CLOSED, HTTP_CONTEXT_STATE_CLOSED)) {
httpRemoveContextFromEpoll(pContext);
httpDebug("context:%p, fd:%d, ip:%s, last state:ready, keepAlive:true, close connect",
pContext, pContext->fd, pContext->ipstr);
httpDebug("context:%p, fd:%d, ip:%s, last state:ready, keepAlive:true, close connect", pContext, pContext->fd,
pContext->ipstr);
} else {
httpRemoveContextFromEpoll(pContext);
httpError("context:%p, fd:%d, ip:%s, last state:%s:%d, keepAlive:true, close connect",
pContext, pContext->fd, pContext->ipstr, httpContextStateStr(pContext->state), pContext->state);
httpError("context:%p, fd:%d, ip:%s, last state:%s:%d, keepAlive:true, close connect", pContext, pContext->fd,
pContext->ipstr, httpContextStateStr(pContext->state), pContext->state);
}
} else {
httpRemoveContextFromEpoll(pContext);
httpDebug("context:%p, fd:%d, ip:%s, last state:%s:%d, keepAlive:false, close connect",
pContext, pContext->fd, pContext->ipstr, httpContextStateStr(pContext->state), pContext->state);
httpDebug("context:%p, fd:%d, ip:%s, last state:%s:%d, keepAlive:false, close context", pContext, pContext->fd,
pContext->ipstr, httpContextStateStr(pContext->state), pContext->state);
}
httpReleaseContext(pContext);
......@@ -214,7 +215,7 @@ void httpCloseContextByServer(HttpContext *pContext) {
} else if (httpAlterContextState(pContext, HTTP_CONTEXT_STATE_DROPPING, HTTP_CONTEXT_STATE_DROPPING)) {
httpDebug("context:%p, fd:%d, ip:%s, epoll already finished, wait app finished", pContext, pContext->fd, pContext->ipstr);
} else if (httpAlterContextState(pContext, HTTP_CONTEXT_STATE_READY, HTTP_CONTEXT_STATE_CLOSED)) {
httpDebug("context:%p, fd:%d, ip:%s, epoll finished, close context", pContext, pContext->fd, pContext->ipstr);
httpDebug("context:%p, fd:%d, ip:%s, epoll finished, close connect", pContext, pContext->fd, pContext->ipstr);
} else if (httpAlterContextState(pContext, HTTP_CONTEXT_STATE_CLOSED, HTTP_CONTEXT_STATE_CLOSED)) {
httpDebug("context:%p, fd:%d, ip:%s, epoll finished, will be closed soon", pContext, pContext->fd, pContext->ipstr);
} else {
......
src/plugins/http/src/httpHandle.c
浏览文件 @ 70625808
......@@ -313,9 +313,9 @@ bool httpParseRequest(HttpContext* pContext) {
return true;
}
httpTraceL("context:%p, fd:%d, ip:%s, thread:%s, numOfFds:%d, read size:%d, raw data:\n%s", pContext, pContext->fd,
pContext->ipstr, pContext->pThread->label, pContext->pThread->numOfFds, pContext->parser.bufsize,
pContext->parser.buffer);
httpTraceL("context:%p, fd:%d, ip:%s, thread:%s, numOfContexts:%d, read size:%d, raw data:\n%s", pContext,
pContext->fd, pContext->ipstr, pContext->pThread->label, pContext->pThread->numOfContexts,
pContext->parser.bufsize, pContext->parser.buffer);
if (!httpGetHttpMethod(pContext)) {
return false;
......
src/plugins/http/src/httpServer.c
浏览文件 @ 70625808
......@@ -293,7 +293,7 @@ static void *httpAcceptHttpConnection(void *arg) {
totalFds = 1;
for (int i = 0; i < pServer->numOfThreads; ++i) {
totalFds += pServer->pThreads[i].numOfFds;
totalFds += pServer->pThreads[i].numOfContexts;
}
if (totalFds > tsHttpCacheSessions * 100) {
......@@ -332,9 +332,9 @@ static void *httpAcceptHttpConnection(void *arg) {
}
// notify the data process, add into the FdObj list
atomic_add_fetch_32(&pThread->numOfFds, 1);
httpDebug("context:%p, fd:%d, ip:%s, thread:%s numOfFds:%d totalFds:%d, accept a new connection", pContext, connFd,
pContext->ipstr, pThread->label, pThread->numOfFds, totalFds);
atomic_add_fetch_32(&pThread->numOfContexts, 1);
httpDebug("context:%p, fd:%d, ip:%s, thread:%s numOfContexts:%d totalFds:%d, accept a new connection", pContext,
connFd, pContext->ipstr, pThread->label, pThread->numOfContexts, totalFds);
// pick up next thread for next connection
threadId++;
......
src/plugins/http/src/httpSql.c
浏览文件 @ 70625808
......@@ -233,10 +233,11 @@ void httpProcessSingleSqlRetrieveCallBack(void *param, TAOS_RES *result, int num
}
}
void httpProcessSingleSqlCallBack(void *param, TAOS_RES *result, int code) {
void httpProcessSingleSqlCallBack(void *param, TAOS_RES *result, int unUsedCode) {
HttpContext *pContext = (HttpContext *)param;
if (pContext == NULL) return;
int32_t code = taos_errno(result);
HttpEncodeMethod *encode = pContext->encodeMethod;
if (code == TSDB_CODE_TSC_ACTION_IN_PROGRESS) {
......@@ -260,8 +261,8 @@ void httpProcessSingleSqlCallBack(void *param, TAOS_RES *result, int code) {
return;
}
int num_fields = taos_field_count(result);
if (num_fields == 0) {
bool isUpdate = tscIsUpdateQuery(result);
if (isUpdate) {
// not select or show commands
int affectRows = taos_affected_rows(result);
......
src/query/src/qExecutor.c
浏览文件 @ 70625808
......@@ -4475,6 +4475,8 @@ static void sequentialTableProcess(SQInfo *pQInfo) {
}
pRuntimeEnv->pQueryHandle = tsdbQueryTables(pQInfo->tsdb, &cond, &gp, pQInfo);
taosArrayDestroy(g1);
taosArrayDestroy(tx);
SArray* s = tsdbGetQueriedTableList(pRuntimeEnv->pQueryHandle);
assert(taosArrayGetSize(s) >= 1);
......@@ -5857,6 +5859,18 @@ _error:
return code;
}
static void freeColumnFilterInfo(SColumnFilterInfo* pFilter, int32_t numOfFilters) {
if (pFilter == NULL) {
return;
}
for (int32_t i = 0; i < numOfFilters; i++) {
if (pFilter[i].filterstr) {
free((void*)(pFilter[i].pz));
}
}
free(pFilter);
}
static void freeQInfo(SQInfo *pQInfo) {
if (!isValidQInfo(pQInfo)) {
return;
......@@ -5925,7 +5939,15 @@ static void freeQInfo(SQInfo *pQInfo) {
tfree(pQuery->tagColList);
tfree(pQuery->pFilterInfo);
tfree(pQuery->colList);
if (pQuery->colList != NULL) {
for (int32_t i = 0; i < pQuery->numOfCols; i++) {
SColumnInfo* column = pQuery->colList + i;
freeColumnFilterInfo(column->filters, column->numOfFilters);
}
tfree(pQuery->colList);
}
tfree(pQuery->sdata);
tfree(pQuery);
......@@ -6122,6 +6144,11 @@ _over:
free(pExprMsg);
taosArrayDestroy(pTableIdList);
for (int32_t i = 0; i < pQueryMsg->numOfCols; i++) {
SColumnInfo* column = pQueryMsg->colList + i;
freeColumnFilterInfo(column->filters, column->numOfFilters);
}
//pQInfo already freed in initQInfo, but *pQInfo may not pointer to null;
if (code != TSDB_CODE_SUCCESS) {
*pQInfo = NULL;
......
src/tsdb/src/tsdbMemTable.c
浏览文件 @ 70625808
......@@ -119,7 +119,8 @@ int tsdbInsertRowToMem(STsdbRepo *pRepo, SDataRow row, STable *pTable) {
int tsdbRefMemTable(STsdbRepo *pRepo, SMemTable *pMemTable) {
if (pMemTable == NULL) return 0;
T_REF_INC(pMemTable);
int ref = T_REF_INC(pMemTable);
tsdbDebug("vgId:%d ref memtable %p ref %d", REPO_ID(pRepo), pMemTable, ref);
return 0;
}
......@@ -127,7 +128,9 @@ int tsdbRefMemTable(STsdbRepo *pRepo, SMemTable *pMemTable) {
int tsdbUnRefMemTable(STsdbRepo *pRepo, SMemTable *pMemTable) {
if (pMemTable == NULL) return 0;
if (T_REF_DEC(pMemTable) == 0) {
int ref = T_REF_DEC(pMemTable);
tsdbDebug("vgId:%d unref memtable %p ref %d", REPO_ID(pRepo), pMemTable, ref);
if (ref == 0) {
STsdbCfg * pCfg = &pRepo->config;
STsdbBufPool *pBufPool = pRepo->pPool;
......@@ -167,6 +170,7 @@ int tsdbTakeMemSnapshot(STsdbRepo *pRepo, SMemTable **pMem, SMemTable **pIMem) {
tsdbRefMemTable(pRepo, *pIMem);
if (tsdbUnlockRepo(pRepo) < 0) return -1;
tsdbDebug("vgId:%d take memory snapshot, pMem %p pIMem %p", REPO_ID(pRepo), *pMem, *pIMem);
return 0;
}
......
src/util/inc/tfile.h
浏览文件 @ 70625808
......@@ -18,6 +18,7 @@
#ifdef TAOS_RANDOM_FILE_FAIL
void taosSetRandomFileFailFactor(int factor);
ssize_t taos_tread(int fd, void *buf, size_t count);
ssize_t taos_twrite(int fd, void *buf, size_t count);
off_t taos_lseek(int fd, off_t offset, int whence);
......
src/util/src/tfile.c
浏览文件 @ 70625808
......@@ -26,40 +26,51 @@
#include "os.h"
#define RANDOM_FILE_FAIL_FACTOR 5
#ifdef TAOS_RANDOM_FILE_FAIL
static int random_file_fail_factor = 20;
void taosSetRandomFileFailFactor(int factor)
{
random_file_fail_factor = factor;
}
#endif
ssize_t taos_tread(int fd, void *buf, size_t count)
{
#ifdef TAOS_RANDOM_FILE_FAIL
if (rand() % RANDOM_FILE_FAIL_FACTOR == 0) {
errno = EIO;
return -1;
if (random_file_fail_factor > 0) {
if (rand() % random_file_fail_factor == 0) {
errno = EIO;
return -1;
}
}
#endif
return tread(fd, buf, count);
}
ssize_t taos_twrite(int fd, void *buf, size_t count)
{
#ifdef TAOS_RANDOM_FILE_FAIL
if (rand() % RANDOM_FILE_FAIL_FACTOR == 0) {
errno = EIO;
return -1;
if (random_file_fail_factor > 0) {
if (rand() % random_file_fail_factor == 0) {
errno = EIO;
return -1;
}
}
#endif
return twrite(fd, buf, count);
}
off_t taos_lseek(int fd, off_t offset, int whence)
{
#ifdef TAOS_RANDOM_FILE_FAIL
if (rand() % RANDOM_FILE_FAIL_FACTOR == 0) {
errno = EIO;
return -1;
if (random_file_fail_factor > 0) {
if (rand() % random_file_fail_factor == 0) {
errno = EIO;
return -1;
}
}
#endif
return lseek(fd, offset, whence);
}
src/wal/src/walMain.c
浏览文件 @ 70625808
......@@ -28,6 +28,7 @@
#include "taoserror.h"
#include "twal.h"
#include "tqueue.h"
#include "tfile.h"
#define walPrefix "wal"
......@@ -180,7 +181,7 @@ int walWrite(void *handle, SWalHead *pHead) {
taosCalcChecksumAppend(0, (uint8_t *)pHead, sizeof(SWalHead));
int contLen = pHead->len + sizeof(SWalHead);
if(write(pWal->fd, pHead, contLen) != contLen) {
if(twrite(pWal->fd, pHead, contLen) != contLen) {
wError("wal:%s, failed to write(%s)", pWal->name, strerror(errno));
terrno = TAOS_SYSTEM_ERROR(errno);
} else {
......@@ -325,7 +326,7 @@ static int walRestoreWalFile(SWal *pWal, void *pVnode, FWalWrite writeFp) {
wDebug("wal:%s, start to restore", name);
while (1) {
int ret = read(fd, pHead, sizeof(SWalHead));
int ret = tread(fd, pHead, sizeof(SWalHead));
if ( ret == 0) break;
if (ret != sizeof(SWalHead)) {
......@@ -340,7 +341,7 @@ static int walRestoreWalFile(SWal *pWal, void *pVnode, FWalWrite writeFp) {
break;
}
ret = read(fd, pHead->cont, pHead->len);
ret = tread(fd, pHead->cont, pHead->len);
if ( ret != pHead->len) {
wWarn("wal:%s, failed to read body, skip, len:%d ret:%d", name, pHead->len, ret);
terrno = TAOS_SYSTEM_ERROR(errno);
......
tests/pytest/crash_gen.py
浏览文件 @ 70625808
#-----!/usr/bin/python3.7
# -----!/usr/bin/python3.7
###################################################################
# Copyright (c) 2016 by TAOS Technologies, Inc.
# All rights reserved.
......@@ -11,7 +11,31 @@
###################################################################
# -*- coding: utf-8 -*-
from __future__ import annotations # For type hinting before definition, ref: https://stackoverflow.com/questions/33533148/how-do-i-specify-that-the-return-type-of-a-method-is-the-same-as-the-class-itsel
# For type hinting before definition, ref:
# https://stackoverflow.com/questions/33533148/how-do-i-specify-that-the-return-type-of-a-method-is-the-same-as-the-class-itsel
from __future__ import annotations
import taos
import crash_gen
from util.sql import *
from util.cases import *
from util.dnodes import *
from util.log import *
from queue import Queue, Empty
from typing import IO
from typing import Set
from typing import Dict
from typing import List
from requests.auth import HTTPBasicAuth
import textwrap
import datetime
import logging
import time
import random
import threading
import requests
import copy
import argparse
import getopt
import sys
import os
......@@ -22,71 +46,48 @@ import traceback
if sys.version_info[0] < 3:
raise Exception("Must be using Python 3")
import getopt
import argparse
import copy
import requests
import threading
import random
import time
import logging
import datetime
import textwrap
import requests
from requests.auth import HTTPBasicAuth
from typing import List
from typing import Dict
from typing import Set
from typing import IO
from queue import Queue, Empty
from util.log import *
from util.dnodes import *
from util.cases import *
from util.sql import *
import crash_gen
import taos
# Global variables, tried to keep a small number.
# Global variables, tried to keep a small number.
# Command-line/Environment Configurations, will set a bit later
# ConfigNameSpace = argparse.Namespace
gConfig = argparse.Namespace() # Dummy value, will be replaced later
gConfig = argparse.Namespace() # Dummy value, will be replaced later
logger = None
def runThread(wt: WorkerThread):
def runThread(wt: WorkerThread):
wt.run()
class CrashGenError(Exception):
def __init__(self, msg=None, errno=None):
self.msg = msg
self.msg = msg
self.errno = errno
def __str__(self):
return self.msg
class WorkerThread:
def __init__(self, pool: ThreadPool, tid,
tc: ThreadCoordinator,
# te: TaskExecutor,
): # note: main thread context!
# self._curStep = -1
def __init__(self, pool: ThreadPool, tid,
tc: ThreadCoordinator,
# te: TaskExecutor,
): # note: main thread context!
# self._curStep = -1
self._pool = pool
self._tid = tid
self._tc = tc # type: ThreadCoordinator
self._tid = tid
self._tc = tc # type: ThreadCoordinator
# self.threadIdent = threading.get_ident()
self._thread = threading.Thread(target=runThread, args=(self,))
self._stepGate = threading.Event()
# Let us have a DB connection of our own
if ( gConfig.per_thread_db_connection ): # type: ignore
if (gConfig.per_thread_db_connection): # type: ignore
# print("connector_type = {}".format(gConfig.connector_type))
self._dbConn = DbConn.createNative() if (gConfig.connector_type == 'native') else DbConn.createRest()
self._dbConn = DbConn.createNative() if (
gConfig.connector_type == 'native') else DbConn.createRest()
self._dbInUse = False # if "use db" was executed already
self._dbInUse = False # if "use db" was executed already
def logDebug(self, msg):
logger.debug(" TRD[{}] {}".format(self._tid, msg))
......@@ -98,137 +99,153 @@ class WorkerThread:
return self._dbInUse
def useDb(self):
if ( not self._dbInUse ):
if (not self._dbInUse):
self.execSql("use db")
self._dbInUse = True
def getTaskExecutor(self):
return self._tc.getTaskExecutor()
return self._tc.getTaskExecutor()
def start(self):
self._thread.start() # AFTER the thread is recorded
def run(self):
def run(self):
# initialization after thread starts, in the thread context
# self.isSleeping = False
logger.info("Starting to run thread: {}".format(self._tid))
if ( gConfig.per_thread_db_connection ): # type: ignore
if (gConfig.per_thread_db_connection): # type: ignore
logger.debug("Worker thread openning database connection")
self._dbConn.open()
self._doTaskLoop()
self._doTaskLoop()
# clean up
if ( gConfig.per_thread_db_connection ): # type: ignore
if (gConfig.per_thread_db_connection): # type: ignore
self._dbConn.close()
def _doTaskLoop(self) :
def _doTaskLoop(self):
# while self._curStep < self._pool.maxSteps:
# tc = ThreadCoordinator(None)
while True:
tc = self._tc # Thread Coordinator, the overall master
while True:
tc = self._tc # Thread Coordinator, the overall master
tc.crossStepBarrier() # shared barrier first, INCLUDING the last one
logger.debug("[TRD] Worker thread [{}] exited barrier...".format(self._tid))
logger.debug(
"[TRD] Worker thread [{}] exited barrier...".format(
self._tid))
self.crossStepGate() # then per-thread gate, after being tapped
logger.debug("[TRD] Worker thread [{}] exited step gate...".format(self._tid))
logger.debug(
"[TRD] Worker thread [{}] exited step gate...".format(
self._tid))
if not self._tc.isRunning():
logger.debug("[TRD] Thread Coordinator not running any more, worker thread now stopping...")
logger.debug(
"[TRD] Thread Coordinator not running any more, worker thread now stopping...")
break
# Fetch a task from the Thread Coordinator
logger.debug("[TRD] Worker thread [{}] about to fetch task".format(self._tid))
logger.debug(
"[TRD] Worker thread [{}] about to fetch task".format(
self._tid))
task = tc.fetchTask()
# Execute such a task
logger.debug("[TRD] Worker thread [{}] about to execute task: {}".format(self._tid, task.__class__.__name__))
logger.debug(
"[TRD] Worker thread [{}] about to execute task: {}".format(
self._tid, task.__class__.__name__))
task.execute(self)
tc.saveExecutedTask(task)
logger.debug("[TRD] Worker thread [{}] finished executing task".format(self._tid))
logger.debug(
"[TRD] Worker thread [{}] finished executing task".format(
self._tid))
self._dbInUse = False # there may be changes between steps
self._dbInUse = False # there may be changes between steps
def verifyThreadSelf(self): # ensure we are called by this own thread
if ( threading.get_ident() != self._thread.ident ):
def verifyThreadSelf(self): # ensure we are called by this own thread
if (threading.get_ident() != self._thread.ident):
raise RuntimeError("Unexpectly called from other threads")
def verifyThreadMain(self): # ensure we are called by the main thread
if ( threading.get_ident() != threading.main_thread().ident ):
def verifyThreadMain(self): # ensure we are called by the main thread
if (threading.get_ident() != threading.main_thread().ident):
raise RuntimeError("Unexpectly called from other threads")
def verifyThreadAlive(self):
if ( not self._thread.is_alive() ):
if (not self._thread.is_alive()):
raise RuntimeError("Unexpected dead thread")
# A gate is different from a barrier in that a thread needs to be "tapped"
def crossStepGate(self):
self.verifyThreadAlive()
self.verifyThreadSelf() # only allowed by ourselves
self.verifyThreadSelf() # only allowed by ourselves
# Wait again at the "gate", waiting to be "tapped"
logger.debug("[TRD] Worker thread {} about to cross the step gate".format(self._tid))
self._stepGate.wait()
logger.debug(
"[TRD] Worker thread {} about to cross the step gate".format(
self._tid))
self._stepGate.wait()
self._stepGate.clear()
# self._curStep += 1 # off to a new step...
def tapStepGate(self): # give it a tap, release the thread waiting there
def tapStepGate(self): # give it a tap, release the thread waiting there
self.verifyThreadAlive()
self.verifyThreadMain() # only allowed for main thread
self.verifyThreadMain() # only allowed for main thread
logger.debug("[TRD] Tapping worker thread {}".format(self._tid))
self._stepGate.set() # wake up!
time.sleep(0) # let the released thread run a bit
self._stepGate.set() # wake up!
time.sleep(0) # let the released thread run a bit
def execSql(self, sql): # TODO: expose DbConn directly
if ( gConfig.per_thread_db_connection ):
return self._dbConn.execute(sql)
def execSql(self, sql): # TODO: expose DbConn directly
if (gConfig.per_thread_db_connection):
return self._dbConn.execute(sql)
else:
return self._tc.getDbManager().getDbConn().execute(sql)
def querySql(self, sql): # TODO: expose DbConn directly
if ( gConfig.per_thread_db_connection ):
return self._dbConn.query(sql)
def querySql(self, sql): # TODO: expose DbConn directly
if (gConfig.per_thread_db_connection):
return self._dbConn.query(sql)
else:
return self._tc.getDbManager().getDbConn().query(sql)
def getQueryResult(self):
if ( gConfig.per_thread_db_connection ):
return self._dbConn.getQueryResult()
if (gConfig.per_thread_db_connection):
return self._dbConn.getQueryResult()
else:
return self._tc.getDbManager().getDbConn().getQueryResult()
def getDbConn(self):
if ( gConfig.per_thread_db_connection ):
return self._dbConn
if (gConfig.per_thread_db_connection):
return self._dbConn
else:
return self._tc.getDbManager().getDbConn()
# def querySql(self, sql): # not "execute", since we are out side the DB context
# if ( gConfig.per_thread_db_connection ):
# return self._dbConn.query(sql)
# return self._dbConn.query(sql)
# else:
# return self._tc.getDbState().getDbConn().query(sql)
# The coordinator of all worker threads, mostly running in main thread
class ThreadCoordinator:
def __init__(self, pool: ThreadPool, dbManager):
self._curStep = -1 # first step is 0
self._curStep = -1 # first step is 0
self._pool = pool
# self._wd = wd
self._te = None # prepare for every new step
self._te = None # prepare for every new step
self._dbManager = dbManager
self._executedTasks: List[Task] = [] # in a given step
self._lock = threading.RLock() # sync access for a few things
self._executedTasks: List[Task] = [] # in a given step
self._lock = threading.RLock() # sync access for a few things
self._stepBarrier = threading.Barrier(self._pool.numThreads + 1) # one barrier for all threads
self._stepBarrier = threading.Barrier(
self._pool.numThreads + 1) # one barrier for all threads
self._execStats = ExecutionStats()
self._runStatus = MainExec.STATUS_RUNNING
def getTaskExecutor(self):
return self._te
def getDbManager(self) -> DbManager :
def getDbManager(self) -> DbManager:
return self._dbManager
def crossStepBarrier(self):
......@@ -238,89 +255,103 @@ class ThreadCoordinator:
self._runStatus = MainExec.STATUS_STOPPING
self._execStats.registerFailure("User Interruption")
def run(self):
def run(self):
self._pool.createAndStartThreads(self)
# Coordinate all threads step by step
self._curStep = -1 # not started yet
maxSteps = gConfig.max_steps # type: ignore
self._execStats.startExec() # start the stop watch
self._curStep = -1 # not started yet
maxSteps = gConfig.max_steps # type: ignore
self._execStats.startExec() # start the stop watch
transitionFailed = False
hasAbortedTask = False
while(self._curStep < maxSteps-1 and
(not transitionFailed) and
(self._runStatus==MainExec.STATUS_RUNNING) and
(not hasAbortedTask)): # maxStep==10, last curStep should be 9
if not gConfig.debug:
print(".", end="", flush=True) # print this only if we are not in debug mode
while(self._curStep < maxSteps - 1 and
(not transitionFailed) and
(self._runStatus == MainExec.STATUS_RUNNING) and
(not hasAbortedTask)): # maxStep==10, last curStep should be 9
if not gConfig.debug:
# print this only if we are not in debug mode
print(".", end="", flush=True)
logger.debug("[TRD] Main thread going to sleep")
# Now main thread (that's us) is ready to enter a step
self.crossStepBarrier() # let other threads go past the pool barrier, but wait at the thread gate
self._stepBarrier.reset() # Other worker threads should now be at the "gate"
# let other threads go past the pool barrier, but wait at the
# thread gate
self.crossStepBarrier()
self._stepBarrier.reset() # Other worker threads should now be at the "gate"
# At this point, all threads should be pass the overall "barrier" and before the per-thread "gate"
# We use this period to do house keeping work, when all worker threads are QUIET.
# We use this period to do house keeping work, when all worker
# threads are QUIET.
hasAbortedTask = False
for task in self._executedTasks :
if task.isAborted() :
for task in self._executedTasks:
if task.isAborted():
print("Task aborted: {}".format(task))
hasAbortedTask = True
break
if hasAbortedTask : # do transition only if tasks are error free
if hasAbortedTask: # do transition only if tasks are error free
self._execStats.registerFailure("Aborted Task Encountered")
else:
else:
try:
sm = self._dbManager.getStateMachine()
logger.debug("[STT] starting transitions")
sm.transition(self._executedTasks) # at end of step, transiton the DB state
# at end of step, transiton the DB state
sm.transition(self._executedTasks)
logger.debug("[STT] transition ended")
# Due to limitation (or maybe not) of the Python library, we cannot share connections across threads
if sm.hasDatabase() :
# Due to limitation (or maybe not) of the Python library,
# we cannot share connections across threads
if sm.hasDatabase():
for t in self._pool.threadList:
logger.debug("[DB] use db for all worker threads")
t.useDb()
# t.execSql("use db") # main thread executing "use db" on behalf of every worker thread
# t.execSql("use db") # main thread executing "use
# db" on behalf of every worker thread
except taos.error.ProgrammingError as err:
if ( err.msg == 'network unavailable' ): # broken DB connection
if (err.msg == 'network unavailable'): # broken DB connection
logger.info("DB connection broken, execution failed")
traceback.print_stack()
transitionFailed = True
self._te = None # Not running any more
self._te = None # Not running any more
self._execStats.registerFailure("Broken DB Connection")
# continue # don't do that, need to tap all threads at end, and maybe signal them to stop
# continue # don't do that, need to tap all threads at
# end, and maybe signal them to stop
else:
raise
raise
# finally:
# pass
self.resetExecutedTasks() # clear the tasks after we are done
self.resetExecutedTasks() # clear the tasks after we are done
# Get ready for next step
logger.debug("<-- Step {} finished".format(self._curStep))
self._curStep += 1 # we are about to get into next step. TODO: race condition here!
logger.debug("\r\n\n--> Step {} starts with main thread waking up".format(self._curStep)) # Now not all threads had time to go to sleep
self._curStep += 1 # we are about to get into next step. TODO: race condition here!
# Now not all threads had time to go to sleep
logger.debug(
"\r\n\n--> Step {} starts with main thread waking up".format(self._curStep))
# A new TE for the new step
if not transitionFailed: # only if not failed
if not transitionFailed: # only if not failed
self._te = TaskExecutor(self._curStep)
logger.debug("[TRD] Main thread waking up at step {}, tapping worker threads".format(self._curStep)) # Now not all threads had time to go to sleep
self.tapAllThreads() # Worker threads will wake up at this point, and each execute it's own task
logger.debug(
"[TRD] Main thread waking up at step {}, tapping worker threads".format(
self._curStep)) # Now not all threads had time to go to sleep
# Worker threads will wake up at this point, and each execute it's
# own task
self.tapAllThreads()
logger.debug("Main thread ready to finish up...")
if not transitionFailed: # only in regular situations
self.crossStepBarrier() # Cross it one last time, after all threads finish
if not transitionFailed: # only in regular situations
self.crossStepBarrier() # Cross it one last time, after all threads finish
self._stepBarrier.reset()
logger.debug("Main thread in exclusive zone...")
self._te = None # No more executor, time to end
self._te = None # No more executor, time to end
logger.debug("Main thread tapping all threads one last time...")
self.tapAllThreads() # Let the threads run one last time
self.tapAllThreads() # Let the threads run one last time
logger.debug("Main thread joining all threads")
self._pool.joinAll() # Get all threads to finish
self._pool.joinAll() # Get all threads to finish
logger.info("\nAll worker threads finished")
self._execStats.endExec()
......@@ -333,24 +364,27 @@ class ThreadCoordinator:
def getExecStats(self):
return self._execStats
def tapAllThreads(self): # in a deterministic manner
def tapAllThreads(self): # in a deterministic manner
wakeSeq = []
for i in range(self._pool.numThreads): # generate a random sequence
if Dice.throw(2) == 1 :
for i in range(self._pool.numThreads): # generate a random sequence
if Dice.throw(2) == 1:
wakeSeq.append(i)
else:
wakeSeq.insert(0, i)
logger.debug("[TRD] Main thread waking up worker threads: {}".format(str(wakeSeq)))
logger.debug(
"[TRD] Main thread waking up worker threads: {}".format(
str(wakeSeq)))
# TODO: set dice seed to a deterministic value
for i in wakeSeq:
self._pool.threadList[i].tapStepGate() # TODO: maybe a bit too deep?!
time.sleep(0) # yield
# TODO: maybe a bit too deep?!
self._pool.threadList[i].tapStepGate()
time.sleep(0) # yield
def isRunning(self):
return self._te != None
return self._te is not None
def fetchTask(self) -> Task :
if ( not self.isRunning() ): # no task
def fetchTask(self) -> Task:
if (not self.isRunning()): # no task
raise RuntimeError("Cannot fetch task when not running")
# return self._wd.pickTask()
# Alternatively, let's ask the DbState for the appropriate task
......@@ -361,31 +395,36 @@ class ThreadCoordinator:
# logger.debug(" (dice:{}/{}) ".format(i, nTasks))
# # return copy.copy(tasks[i]) # Needs a fresh copy, to save execution results, etc.
# return tasks[i].clone() # TODO: still necessary?
taskType = self.getDbManager().getStateMachine().pickTaskType() # pick a task type for current state
return taskType(self.getDbManager(), self._execStats) # create a task from it
# pick a task type for current state
taskType = self.getDbManager().getStateMachine().pickTaskType()
return taskType(
self.getDbManager(),
self._execStats) # create a task from it
def resetExecutedTasks(self):
self._executedTasks = [] # should be under single thread
self._executedTasks = [] # should be under single thread
def saveExecutedTask(self, task):
with self._lock:
self._executedTasks.append(task)
# We define a class to run a number of threads in locking steps.
class ThreadPool:
def __init__(self, numThreads, maxSteps):
self.numThreads = numThreads
self.maxSteps = maxSteps
# Internal class variables
self.curStep = 0
self.threadList = [] # type: List[WorkerThread]
self.threadList = [] # type: List[WorkerThread]
# starting to run all the threads, in locking steps
def createAndStartThreads(self, tc: ThreadCoordinator):
for tid in range(0, self.numThreads): # Create the threads
workerThread = WorkerThread(self, tid, tc)
for tid in range(0, self.numThreads): # Create the threads
workerThread = WorkerThread(self, tid, tc)
self.threadList.append(workerThread)
workerThread.start() # start, but should block immediately before step 0
workerThread.start() # start, but should block immediately before step 0
def joinAll(self):
for workerThread in self.threadList:
......@@ -394,21 +433,24 @@ class ThreadPool:
# A queue of continguous POSITIVE integers, used by DbManager to generate continuous numbers
# for new table names
class LinearQueue():
def __init__(self):
self.firstIndex = 1 # 1st ever element
self.lastIndex = 0
self._lock = threading.RLock() # our functions may call each other
self.inUse = set() # the indexes that are in use right now
self._lock = threading.RLock() # our functions may call each other
self.inUse = set() # the indexes that are in use right now
def toText(self):
return "[{}..{}], in use: {}".format(self.firstIndex, self.lastIndex, self.inUse)
return "[{}..{}], in use: {}".format(
self.firstIndex, self.lastIndex, self.inUse)
# Push (add new element, largest) to the tail, and mark it in use
def push(self):
def push(self):
with self._lock:
# if ( self.isEmpty() ):
# self.lastIndex = self.firstIndex
# if ( self.isEmpty() ):
# self.lastIndex = self.firstIndex
# return self.firstIndex
# Otherwise we have something
self.lastIndex += 1
......@@ -418,12 +460,12 @@ class LinearQueue():
def pop(self):
with self._lock:
if ( self.isEmpty() ):
# raise RuntimeError("Cannot pop an empty queue")
return False # TODO: None?
if (self.isEmpty()):
# raise RuntimeError("Cannot pop an empty queue")
return False # TODO: None?
index = self.firstIndex
if ( index in self.inUse ):
if (index in self.inUse):
return False
self.firstIndex += 1
......@@ -441,33 +483,35 @@ class LinearQueue():
def allocate(self, i):
with self._lock:
# logger.debug("LQ allocating item {}".format(i))
if ( i in self.inUse ):
raise RuntimeError("Cannot re-use same index in queue: {}".format(i))
if (i in self.inUse):
raise RuntimeError(
"Cannot re-use same index in queue: {}".format(i))
self.inUse.add(i)
def release(self, i):
with self._lock:
# logger.debug("LQ releasing item {}".format(i))
self.inUse.remove(i) # KeyError possible, TODO: why?
self.inUse.remove(i) # KeyError possible, TODO: why?
def size(self):
return self.lastIndex + 1 - self.firstIndex
def pickAndAllocate(self):
if ( self.isEmpty() ):
if (self.isEmpty()):
return None
with self._lock:
cnt = 0 # counting the interations
cnt = 0 # counting the interations
while True:
cnt += 1
if ( cnt > self.size()*10 ): # 10x iteration already
if (cnt > self.size() * 10): # 10x iteration already
# raise RuntimeError("Failed to allocate LinearQueue element")
return None
ret = Dice.throwRange(self.firstIndex, self.lastIndex+1)
if ( not ret in self.inUse ):
ret = Dice.throwRange(self.firstIndex, self.lastIndex + 1)
if (ret not in self.inUse):
self.allocate(ret)
return ret
class DbConn:
TYPE_NATIVE = "native-c"
TYPE_REST = "rest-api"
......@@ -480,7 +524,8 @@ class DbConn:
elif connType == cls.TYPE_REST:
return DbConnRest()
else:
raise RuntimeError("Unexpected connection type: {}".format(connType))
raise RuntimeError(
"Unexpected connection type: {}".format(connType))
@classmethod
def createNative(cls):
......@@ -495,18 +540,21 @@ class DbConn:
self._type = self.TYPE_INVALID
def open(self):
if ( self.isOpen ):
if (self.isOpen):
raise RuntimeError("Cannot re-open an existing DB connection")
# below implemented by child classes
self.openByType()
logger.debug("[DB] data connection opened, type = {}".format(self._type))
logger.debug(
"[DB] data connection opened, type = {}".format(
self._type))
self.isOpen = True
def resetDb(self): # reset the whole database, etc.
if ( not self.isOpen ):
raise RuntimeError("Cannot reset database until connection is open")
def resetDb(self): # reset the whole database, etc.
if (not self.isOpen):
raise RuntimeError(
"Cannot reset database until connection is open")
# self._tdSql.prepare() # Recreate database, etc.
self.execute('drop database if exists db')
......@@ -515,83 +563,99 @@ class DbConn:
# self._cursor.execute('use db')
# tdSql.execute('show databases')
def queryScalar(self, sql) -> int :
def queryScalar(self, sql) -> int:
return self._queryAny(sql)
def queryString(self, sql) -> str :
def queryString(self, sql) -> str:
return self._queryAny(sql)
def _queryAny(self, sql) : # actual query result as an int
if ( not self.isOpen ):
raise RuntimeError("Cannot query database until connection is open")
def _queryAny(self, sql): # actual query result as an int
if (not self.isOpen):
raise RuntimeError(
"Cannot query database until connection is open")
nRows = self.query(sql)
if nRows != 1 :
raise RuntimeError("Unexpected result for query: {}, rows = {}".format(sql, nRows))
if nRows != 1:
raise RuntimeError(
"Unexpected result for query: {}, rows = {}".format(
sql, nRows))
if self.getResultRows() != 1 or self.getResultCols() != 1:
raise RuntimeError("Unexpected result set for query: {}".format(sql))
raise RuntimeError(
"Unexpected result set for query: {}".format(sql))
return self.getQueryResult()[0][0]
def execute(self, sql):
raise RuntimeError("Unexpected execution, should be overriden")
def openByType(self):
raise RuntimeError("Unexpected execution, should be overriden")
def getQueryResult(self):
raise RuntimeError("Unexpected execution, should be overriden")
def getResultRows(self):
raise RuntimeError("Unexpected execution, should be overriden")
def getResultCols(self):
raise RuntimeError("Unexpected execution, should be overriden")
# Sample: curl -u root:taosdata -d "show databases" localhost:6020/rest/sql
class DbConnRest(DbConn):
def __init__(self):
super().__init__()
self._type = self.TYPE_REST
self._url = "http://localhost:6020/rest/sql" # fixed for now
self._url = "http://localhost:6020/rest/sql" # fixed for now
self._result = None
def openByType(self): # Open connection
pass # do nothing, always open
def openByType(self): # Open connection
pass # do nothing, always open
def close(self):
if ( not self.isOpen ):
raise RuntimeError("Cannot clean up database until connection is open")
if (not self.isOpen):
raise RuntimeError(
"Cannot clean up database until connection is open")
# Do nothing for REST
logger.debug("[DB] REST Database connection closed")
self.isOpen = False
def _doSql(self, sql):
r = requests.post(self._url,
data = sql,
auth = HTTPBasicAuth('root', 'taosdata'))
r = requests.post(self._url,
data=sql,
auth=HTTPBasicAuth('root', 'taosdata'))
rj = r.json()
# Sanity check for the "Json Result"
if (not 'status' in rj):
if ('status' not in rj):
raise RuntimeError("No status in REST response")
if rj['status'] == 'error': # clearly reported error
if (not 'code' in rj): # error without code
raise RuntimeError("REST error return without code")
errno = rj['code'] # May need to massage this in the future
if rj['status'] == 'error': # clearly reported error
if ('code' not in rj): # error without code
raise RuntimeError("REST error return without code")
errno = rj['code'] # May need to massage this in the future
# print("Raising programming error with REST return: {}".format(rj))
raise taos.error.ProgrammingError(rj['desc'], errno) # todo: check existance of 'desc'
raise taos.error.ProgrammingError(
rj['desc'], errno) # todo: check existance of 'desc'
if rj['status'] != 'succ': # better be this
raise RuntimeError("Unexpected REST return status: {}".format(rj['status']))
if rj['status'] != 'succ': # better be this
raise RuntimeError(
"Unexpected REST return status: {}".format(
rj['status']))
nRows = rj['rows'] if ('rows' in rj) else 0
self._result = rj
self._result = rj
return nRows
def execute(self, sql):
if ( not self.isOpen ):
raise RuntimeError("Cannot execute database commands until connection is open")
def execute(self, sql):
if (not self.isOpen):
raise RuntimeError(
"Cannot execute database commands until connection is open")
logger.debug("[SQL-REST] Executing SQL: {}".format(sql))
nRows = self._doSql(sql)
logger.debug("[SQL-REST] Execution Result, nRows = {}, SQL = {}".format(nRows, sql))
logger.debug(
"[SQL-REST] Execution Result, nRows = {}, SQL = {}".format(nRows, sql))
return nRows
def query(self, sql) : # return rows affected
def query(self, sql): # return rows affected
return self.execute(sql)
def getQueryResult(self):
......@@ -605,48 +669,117 @@ class DbConnRest(DbConn):
def getResultCols(self):
print(self._result)
raise RuntimeError("TBD")
# Duplicate code from TDMySQL, TODO: merge all this into DbConnNative
class MyTDSql:
def __init__(self):
self.queryRows = 0
self.queryCols = 0
self.affectedRows = 0
def init(self, cursor, log=True):
self.cursor = cursor
# if (log):
# caller = inspect.getframeinfo(inspect.stack()[1][0])
# self.cursor.log(caller.filename + ".sql")
def close(self):
self.cursor.close()
def query(self, sql):
self.sql = sql
try:
self.cursor.execute(sql)
self.queryResult = self.cursor.fetchall()
self.queryRows = len(self.queryResult)
self.queryCols = len(self.cursor.description)
except Exception as e:
# caller = inspect.getframeinfo(inspect.stack()[1][0])
# args = (caller.filename, caller.lineno, sql, repr(e))
# tdLog.exit("%s(%d) failed: sql:%s, %s" % args)
raise
return self.queryRows
def execute(self, sql):
self.sql = sql
try:
self.affectedRows = self.cursor.execute(sql)
except Exception as e:
# caller = inspect.getframeinfo(inspect.stack()[1][0])
# args = (caller.filename, caller.lineno, sql, repr(e))
# tdLog.exit("%s(%d) failed: sql:%s, %s" % args)
raise
return self.affectedRows
class DbConnNative(DbConn):
def __init__(self):
super().__init__()
self._type = self.TYPE_REST
self._conn = None
self._conn = None
self._cursor = None
def openByType(self): # Open connection
cfgPath = "../../build/test/cfg"
self._conn = taos.connect(host="127.0.0.1", config=cfgPath) # TODO: make configurable
def getBuildPath(self):
selfPath = os.path.dirname(os.path.realpath(__file__))
if ("community" in selfPath):
projPath = selfPath[:selfPath.find("communit")]
else:
projPath = selfPath[:selfPath.find("tests")]
for root, dirs, files in os.walk(projPath):
if ("taosd" in files):
rootRealPath = os.path.dirname(os.path.realpath(root))
if ("packaging" not in rootRealPath):
buildPath = root[:len(root) - len("/build/bin")]
break
return buildPath
def openByType(self): # Open connection
cfgPath = self.getBuildPath() + "/test/cfg"
self._conn = taos.connect(
host="127.0.0.1",
config=cfgPath) # TODO: make configurable
self._cursor = self._conn.cursor()
# Get the connection/cursor ready
self._cursor.execute('reset query cache')
# self._cursor.execute('use db') # do this at the beginning of every step
# self._cursor.execute('use db') # do this at the beginning of every
# step
# Open connection
self._tdSql = TDSql()
self._tdSql = MyTDSql()
self._tdSql.init(self._cursor)
def close(self):
if ( not self.isOpen ):
raise RuntimeError("Cannot clean up database until connection is open")
if (not self.isOpen):
raise RuntimeError(
"Cannot clean up database until connection is open")
self._tdSql.close()
logger.debug("[DB] Database connection closed")
self.isOpen = False
def execute(self, sql):
if ( not self.isOpen ):
raise RuntimeError("Cannot execute database commands until connection is open")
def execute(self, sql):
if (not self.isOpen):
raise RuntimeError(
"Cannot execute database commands until connection is open")
logger.debug("[SQL] Executing SQL: {}".format(sql))
nRows = self._tdSql.execute(sql)
logger.debug("[SQL] Execution Result, nRows = {}, SQL = {}".format(nRows, sql))
logger.debug(
"[SQL] Execution Result, nRows = {}, SQL = {}".format(
nRows, sql))
return nRows
def query(self, sql) : # return rows affected
if ( not self.isOpen ):
raise RuntimeError("Cannot query database until connection is open")
def query(self, sql): # return rows affected
if (not self.isOpen):
raise RuntimeError(
"Cannot query database until connection is open")
logger.debug("[SQL] Executing SQL: {}".format(sql))
nRows = self._tdSql.query(sql)
logger.debug("[SQL] Query Result, nRows = {}, SQL = {}".format(nRows, sql))
logger.debug(
"[SQL] Query Result, nRows = {}, SQL = {}".format(
nRows, sql))
return nRows
# results are in: return self._tdSql.queryResult
......@@ -659,13 +792,13 @@ class DbConnNative(DbConn):
def getResultCols(self):
return self._tdSql.queryCols
class AnyState:
STATE_INVALID = -1
STATE_EMPTY = 0 # nothing there, no even a DB
STATE_DB_ONLY = 1 # we have a DB, but nothing else
STATE_INVALID = -1
STATE_EMPTY = 0 # nothing there, no even a DB
STATE_DB_ONLY = 1 # we have a DB, but nothing else
STATE_TABLE_ONLY = 2 # we have a table, but totally empty
STATE_HAS_DATA = 3 # we have some data in the table
STATE_HAS_DATA = 3 # we have some data in the table
_stateNames = ["Invalid", "Empty", "DB_Only", "Table_Only", "Has_Data"]
STATE_VAL_IDX = 0
......@@ -680,7 +813,8 @@ class AnyState:
self._info = self.getInfo()
def __str__(self):
return self._stateNames[self._info[self.STATE_VAL_IDX] + 1] # -1 hack to accomodate the STATE_INVALID case
# -1 hack to accomodate the STATE_INVALID case
return self._stateNames[self._info[self.STATE_VAL_IDX] + 1]
def getInfo(self):
raise RuntimeError("Must be overriden by child classes")
......@@ -691,7 +825,9 @@ class AnyState:
elif isinstance(other, AnyState):
return self.getValIndex() == other.getValIndex()
else:
raise RuntimeError("Unexpected comparison, type = {}".format(type(other)))
raise RuntimeError(
"Unexpected comparison, type = {}".format(
type(other)))
def verifyTasksToState(self, tasks, newState):
raise RuntimeError("Must be overriden by child classes")
......@@ -701,55 +837,65 @@ class AnyState:
def getValue(self):
return self._info[self.STATE_VAL_IDX]
def canCreateDb(self):
return self._info[self.CAN_CREATE_DB]
def canDropDb(self):
return self._info[self.CAN_DROP_DB]
def canCreateFixedSuperTable(self):
return self._info[self.CAN_CREATE_FIXED_SUPER_TABLE]
def canDropFixedSuperTable(self):
return self._info[self.CAN_DROP_FIXED_SUPER_TABLE]
def canAddData(self):
return self._info[self.CAN_ADD_DATA]
def canReadData(self):
return self._info[self.CAN_READ_DATA]
def assertAtMostOneSuccess(self, tasks, cls):
sCnt = 0
for task in tasks :
for task in tasks:
if not isinstance(task, cls):
continue
if task.isSuccess():
# task.logDebug("Task success found")
sCnt += 1
if ( sCnt >= 2 ):
raise RuntimeError("Unexpected more than 1 success with task: {}".format(cls))
if (sCnt >= 2):
raise RuntimeError(
"Unexpected more than 1 success with task: {}".format(cls))
def assertIfExistThenSuccess(self, tasks, cls):
sCnt = 0
exists = False
for task in tasks :
for task in tasks:
if not isinstance(task, cls):
continue
exists = True # we have a valid instance
exists = True # we have a valid instance
if task.isSuccess():
sCnt += 1
if ( exists and sCnt <= 0 ):
raise RuntimeError("Unexpected zero success for task: {}".format(cls))
if (exists and sCnt <= 0):
raise RuntimeError(
"Unexpected zero success for task: {}".format(cls))
def assertNoTask(self, tasks, cls):
for task in tasks :
for task in tasks:
if isinstance(task, cls):
raise CrashGenError("This task: {}, is not expected to be present, given the success/failure of others".format(cls.__name__))
raise CrashGenError(
"This task: {}, is not expected to be present, given the success/failure of others".format(cls.__name__))
def assertNoSuccess(self, tasks, cls):
for task in tasks :
for task in tasks:
if isinstance(task, cls):
if task.isSuccess():
raise RuntimeError("Unexpected successful task: {}".format(cls))
raise RuntimeError(
"Unexpected successful task: {}".format(cls))
def hasSuccess(self, tasks, cls):
for task in tasks :
for task in tasks:
if not isinstance(task, cls):
continue
if task.isSuccess():
......@@ -757,35 +903,40 @@ class AnyState:
return False
def hasTask(self, tasks, cls):
for task in tasks :
for task in tasks:
if isinstance(task, cls):
return True
return False
class StateInvalid(AnyState):
def getInfo(self):
return [
self.STATE_INVALID,
False, False, # can create/drop Db
False, False, # can create/drop fixed table
False, False, # can insert/read data with fixed table
False, False, # can create/drop Db
False, False, # can create/drop fixed table
False, False, # can insert/read data with fixed table
]
# def verifyTasksToState(self, tasks, newState):
class StateEmpty(AnyState):
def getInfo(self):
return [
self.STATE_EMPTY,
True, False, # can create/drop Db
False, False, # can create/drop fixed table
False, False, # can insert/read data with fixed table
True, False, # can create/drop Db
False, False, # can create/drop fixed table
False, False, # can insert/read data with fixed table
]
def verifyTasksToState(self, tasks, newState):
if ( self.hasSuccess(tasks, TaskCreateDb) ): # at EMPTY, if there's succes in creating DB
if ( not self.hasTask(tasks, TaskDropDb) ) : # and no drop_db tasks
self.assertAtMostOneSuccess(tasks, TaskCreateDb) # we must have at most one. TODO: compare numbers
def verifyTasksToState(self, tasks, newState):
if (self.hasSuccess(tasks, TaskCreateDb)
): # at EMPTY, if there's succes in creating DB
if (not self.hasTask(tasks, TaskDropDb)): # and no drop_db tasks
# we must have at most one. TODO: compare numbers
self.assertAtMostOneSuccess(tasks, TaskCreateDb)
class StateDbOnly(AnyState):
def getInfo(self):
......@@ -797,32 +948,34 @@ class StateDbOnly(AnyState):
]
def verifyTasksToState(self, tasks, newState):
if ( not self.hasTask(tasks, TaskCreateDb) ):
self.assertAtMostOneSuccess(tasks, TaskDropDb) # only if we don't create any more
if (not self.hasTask(tasks, TaskCreateDb)):
# only if we don't create any more
self.assertAtMostOneSuccess(tasks, TaskDropDb)
self.assertIfExistThenSuccess(tasks, TaskDropDb)
# self.assertAtMostOneSuccess(tasks, CreateFixedTableTask) # not true in massively parrallel cases
# Nothing to be said about adding data task
# if ( self.hasSuccess(tasks, DropDbTask) ): # dropped the DB
# self.assertHasTask(tasks, DropDbTask) # implied by hasSuccess
# self.assertAtMostOneSuccess(tasks, DropDbTask)
# self._state = self.STATE_EMPTY
# self.assertHasTask(tasks, DropDbTask) # implied by hasSuccess
# self.assertAtMostOneSuccess(tasks, DropDbTask)
# self._state = self.STATE_EMPTY
# if ( self.hasSuccess(tasks, TaskCreateSuperTable) ): # did not drop db, create table success
# # self.assertHasTask(tasks, CreateFixedTableTask) # tried to create table
# if ( not self.hasTask(tasks, TaskDropSuperTable) ):
# if ( not self.hasTask(tasks, TaskDropSuperTable) ):
# self.assertAtMostOneSuccess(tasks, TaskCreateSuperTable) # at most 1 attempt is successful, if we don't drop anything
# self.assertNoTask(tasks, DropDbTask) # should have have tried
# if ( not self.hasSuccess(tasks, AddFixedDataTask) ): # just created table, no data yet
# # can't say there's add-data attempts, since they may all fail
# self._state = self.STATE_TABLE_ONLY
# else:
# self._state = self.STATE_HAS_DATA
# self.assertNoTask(tasks, DropDbTask) # should have have tried
# if ( not self.hasSuccess(tasks, AddFixedDataTask) ): # just created table, no data yet
# # can't say there's add-data attempts, since they may all fail
# self._state = self.STATE_TABLE_ONLY
# else:
# self._state = self.STATE_HAS_DATA
# What about AddFixedData?
# elif ( self.hasSuccess(tasks, AddFixedDataTask) ):
# self._state = self.STATE_HAS_DATA
# else: # no success in dropping db tasks, no success in create fixed table? read data should also fail
# # raise RuntimeError("Unexpected no-success scenario") # We might just landed all failure tasks,
# # raise RuntimeError("Unexpected no-success scenario") # We might just landed all failure tasks,
# self._state = self.STATE_DB_ONLY # no change
class StateSuperTableOnly(AnyState):
def getInfo(self):
return [
......@@ -833,9 +986,11 @@ class StateSuperTableOnly(AnyState):
]
def verifyTasksToState(self, tasks, newState):
if ( self.hasSuccess(tasks, TaskDropSuperTable) ): # we are able to drop the table
if (self.hasSuccess(tasks, TaskDropSuperTable)
): # we are able to drop the table
#self.assertAtMostOneSuccess(tasks, TaskDropSuperTable)
self.hasSuccess(tasks, TaskCreateSuperTable) # we must have had recreted it
# we must have had recreted it
self.hasSuccess(tasks, TaskCreateSuperTable)
# self._state = self.STATE_DB_ONLY
# elif ( self.hasSuccess(tasks, AddFixedDataTask) ): # no success dropping the table, but added data
......@@ -849,6 +1004,7 @@ class StateSuperTableOnly(AnyState):
# raise RuntimeError("Unexpected no-success scenarios")
# TODO: need to revamp!!
class StateHasData(AnyState):
def getInfo(self):
return [
......@@ -859,13 +1015,15 @@ class StateHasData(AnyState):
]
def verifyTasksToState(self, tasks, newState):
if ( newState.equals(AnyState.STATE_EMPTY) ):
if (newState.equals(AnyState.STATE_EMPTY)):
self.hasSuccess(tasks, TaskDropDb)
if ( not self.hasTask(tasks, TaskCreateDb) ) :
self.assertAtMostOneSuccess(tasks, TaskDropDb) # TODO: dicy
elif ( newState.equals(AnyState.STATE_DB_ONLY) ): # in DB only
if ( not self.hasTask(tasks, TaskCreateDb)): # without a create_db task
self.assertNoTask(tasks, TaskDropDb) # we must have drop_db task
if (not self.hasTask(tasks, TaskCreateDb)):
self.assertAtMostOneSuccess(tasks, TaskDropDb) # TODO: dicy
elif (newState.equals(AnyState.STATE_DB_ONLY)): # in DB only
if (not self.hasTask(tasks, TaskCreateDb)
): # without a create_db task
# we must have drop_db task
self.assertNoTask(tasks, TaskDropDb)
self.hasSuccess(tasks, TaskDropSuperTable)
# self.assertAtMostOneSuccess(tasks, DropFixedSuperTableTask) # TODO: dicy
# elif ( newState.equals(AnyState.STATE_TABLE_ONLY) ): # data deleted
......@@ -873,19 +1031,26 @@ class StateHasData(AnyState):
# self.assertNoTask(tasks, TaskDropSuperTable)
# self.assertNoTask(tasks, TaskAddData)
# self.hasSuccess(tasks, DeleteDataTasks)
else: # should be STATE_HAS_DATA
if (not self.hasTask(tasks, TaskCreateDb) ): # only if we didn't create one
self.assertNoTask(tasks, TaskDropDb) # we shouldn't have dropped it
if (not self.hasTask(tasks, TaskCreateSuperTable)) : # if we didn't create the table
self.assertNoTask(tasks, TaskDropSuperTable) # we should not have a task that drops it
else: # should be STATE_HAS_DATA
if (not self.hasTask(tasks, TaskCreateDb)
): # only if we didn't create one
# we shouldn't have dropped it
self.assertNoTask(tasks, TaskDropDb)
if (not self.hasTask(tasks, TaskCreateSuperTable)
): # if we didn't create the table
# we should not have a task that drops it
self.assertNoTask(tasks, TaskDropSuperTable)
# self.assertIfExistThenSuccess(tasks, ReadFixedDataTask)
class StateMechine:
def __init__(self, dbConn):
self._dbConn = dbConn
self._curState = self._findCurrentState() # starting state
self._stateWeights = [1,3,5,15] # transitition target probabilities, indexed with value of STATE_EMPTY, STATE_DB_ONLY, etc.
self._curState = self._findCurrentState() # starting state
# transitition target probabilities, indexed with value of STATE_EMPTY,
# STATE_DB_ONLY, etc.
self._stateWeights = [1, 3, 5, 15]
def getCurrentState(self):
return self._curState
......@@ -893,142 +1058,178 @@ class StateMechine:
return self._curState.canDropDb() # ha, can drop DB means it has one
# May be slow, use cautionsly...
def getTaskTypes(self): # those that can run (directly/indirectly) from the current state
def getTaskTypes(self): # those that can run (directly/indirectly) from the current state
def typesToStrings(types):
ss = []
for t in types:
ss.append(t.__name__)
return ss
allTaskClasses = StateTransitionTask.__subclasses__() # all state transition tasks
allTaskClasses = StateTransitionTask.__subclasses__() # all state transition tasks
firstTaskTypes = []
for tc in allTaskClasses:
# t = tc(self) # create task object
# t = tc(self) # create task object
if tc.canBeginFrom(self._curState):
firstTaskTypes.append(tc)
# now we have all the tasks that can begin directly from the current state, let's figure out the INDIRECT ones
taskTypes = firstTaskTypes.copy() # have to have these
for task1 in firstTaskTypes: # each task type gathered so far
endState = task1.getEndState() # figure the end state
if endState == None: # does not change end state
continue # no use, do nothing
for tc in allTaskClasses: # what task can further begin from there?
# now we have all the tasks that can begin directly from the current
# state, let's figure out the INDIRECT ones
taskTypes = firstTaskTypes.copy() # have to have these
for task1 in firstTaskTypes: # each task type gathered so far
endState = task1.getEndState() # figure the end state
if endState is None: # does not change end state
continue # no use, do nothing
for tc in allTaskClasses: # what task can further begin from there?
if tc.canBeginFrom(endState) and (tc not in firstTaskTypes):
taskTypes.append(tc) # gather it
taskTypes.append(tc) # gather it
if len(taskTypes) <= 0:
raise RuntimeError("No suitable task types found for state: {}".format(self._curState))
logger.debug("[OPS] Tasks found for state {}: {}".format(self._curState, typesToStrings(taskTypes)))
raise RuntimeError(
"No suitable task types found for state: {}".format(
self._curState))
logger.debug(
"[OPS] Tasks found for state {}: {}".format(
self._curState,
typesToStrings(taskTypes)))
return taskTypes
def _findCurrentState(self):
dbc = self._dbConn
ts = time.time() # we use this to debug how fast/slow it is to do the various queries to find the current DB state
if dbc.query("show databases") == 0 : # no database?!
ts = time.time() # we use this to debug how fast/slow it is to do the various queries to find the current DB state
if dbc.query("show databases") == 0: # no database?!
# logger.debug("Found EMPTY state")
logger.debug("[STT] empty database found, between {} and {}".format(ts, time.time()))
logger.debug(
"[STT] empty database found, between {} and {}".format(
ts, time.time()))
return StateEmpty()
dbc.execute("use db") # did not do this when openning connection, and this is NOT the worker thread, which does this on their own
if dbc.query("show tables") == 0 : # no tables
# did not do this when openning connection, and this is NOT the worker
# thread, which does this on their own
dbc.execute("use db")
if dbc.query("show tables") == 0: # no tables
# logger.debug("Found DB ONLY state")
logger.debug("[STT] DB_ONLY found, between {} and {}".format(ts, time.time()))
logger.debug(
"[STT] DB_ONLY found, between {} and {}".format(
ts, time.time()))
return StateDbOnly()
if dbc.query("SELECT * FROM db.{}".format(DbManager.getFixedSuperTableName()) ) == 0 : # no regular tables
if dbc.query("SELECT * FROM db.{}".format(DbManager.getFixedSuperTableName())
) == 0: # no regular tables
# logger.debug("Found TABLE_ONLY state")
logger.debug("[STT] SUPER_TABLE_ONLY found, between {} and {}".format(ts, time.time()))
logger.debug(
"[STT] SUPER_TABLE_ONLY found, between {} and {}".format(
ts, time.time()))
return StateSuperTableOnly()
else: # has actual tables
else: # has actual tables
# logger.debug("Found HAS_DATA state")
logger.debug("[STT] HAS_DATA found, between {} and {}".format(ts, time.time()))
logger.debug(
"[STT] HAS_DATA found, between {} and {}".format(
ts, time.time()))
return StateHasData()
def transition(self, tasks):
if ( len(tasks) == 0 ): # before 1st step, or otherwise empty
if (len(tasks) == 0): # before 1st step, or otherwise empty
logger.debug("[STT] Starting State: {}".format(self._curState))
return # do nothing
return # do nothing
self._dbConn.execute("show dnodes") # this should show up in the server log, separating steps
# this should show up in the server log, separating steps
self._dbConn.execute("show dnodes")
# Generic Checks, first based on the start state
if self._curState.canCreateDb():
self._curState.assertIfExistThenSuccess(tasks, TaskCreateDb)
# self.assertAtMostOneSuccess(tasks, CreateDbTask) # not really, in case of multiple creation and drops
# self.assertAtMostOneSuccess(tasks, CreateDbTask) # not really, in
# case of multiple creation and drops
if self._curState.canDropDb():
self._curState.assertIfExistThenSuccess(tasks, TaskDropDb)
# self.assertAtMostOneSuccess(tasks, DropDbTask) # not really in case of drop-create-drop
# self.assertAtMostOneSuccess(tasks, DropDbTask) # not really in
# case of drop-create-drop
# if self._state.canCreateFixedTable():
# self.assertIfExistThenSuccess(tasks, CreateFixedTableTask) # Not true, DB may be dropped
# self.assertAtMostOneSuccess(tasks, CreateFixedTableTask) # not really, in case of create-drop-create
# self.assertAtMostOneSuccess(tasks, CreateFixedTableTask) # not
# really, in case of create-drop-create
# if self._state.canDropFixedTable():
# self.assertIfExistThenSuccess(tasks, DropFixedTableTask) # Not True, the whole DB may be dropped
# self.assertAtMostOneSuccess(tasks, DropFixedTableTask) # not really in case of drop-create-drop
# self.assertAtMostOneSuccess(tasks, DropFixedTableTask) # not
# really in case of drop-create-drop
# if self._state.canAddData():
# self.assertIfExistThenSuccess(tasks, AddFixedDataTask) # not true actually
# self.assertIfExistThenSuccess(tasks, AddFixedDataTask) # not true
# actually
# if self._state.canReadData():
# Nothing for sure
newState = self._findCurrentState()
logger.debug("[STT] New DB state determined: {}".format(newState))
self._curState.verifyTasksToState(tasks, newState) # can old state move to new state through the tasks?
# can old state move to new state through the tasks?
self._curState.verifyTasksToState(tasks, newState)
self._curState = newState
def pickTaskType(self):
taskTypes = self.getTaskTypes() # all the task types we can choose from at curent state
# all the task types we can choose from at curent state
taskTypes = self.getTaskTypes()
weights = []
for tt in taskTypes:
endState = tt.getEndState()
if endState != None :
weights.append(self._stateWeights[endState.getValIndex()]) # TODO: change to a method
if endState is not None:
# TODO: change to a method
weights.append(self._stateWeights[endState.getValIndex()])
else:
weights.append(10) # read data task, default to 10: TODO: change to a constant
# read data task, default to 10: TODO: change to a constant
weights.append(10)
i = self._weighted_choice_sub(weights)
# logger.debug(" (weighted random:{}/{}) ".format(i, len(taskTypes)))
# logger.debug(" (weighted random:{}/{}) ".format(i, len(taskTypes)))
return taskTypes[i]
def _weighted_choice_sub(self, weights): # ref: https://eli.thegreenplace.net/2010/01/22/weighted-random-generation-in-python/
rnd = random.random() * sum(weights) # TODO: use our dice to ensure it being determinstic?
# ref:
# https://eli.thegreenplace.net/2010/01/22/weighted-random-generation-in-python/
def _weighted_choice_sub(self, weights):
# TODO: use our dice to ensure it being determinstic?
rnd = random.random() * sum(weights)
for i, w in enumerate(weights):
rnd -= w
if rnd < 0:
return i
# Manager of the Database Data/Connection
class DbManager():
def __init__(self, resetDb = True):
class DbManager():
def __init__(self, resetDb=True):
self.tableNumQueue = LinearQueue()
self._lastTick = self.setupLastTick() # datetime.datetime(2019, 1, 1) # initial date time tick
self._lastInt = 0 # next one is initial integer
# datetime.datetime(2019, 1, 1) # initial date time tick
self._lastTick = self.setupLastTick()
self._lastInt = 0 # next one is initial integer
self._lock = threading.RLock()
# self.openDbServerConnection()
self._dbConn = DbConn.createNative() if (gConfig.connector_type=='native') else DbConn.createRest()
self._dbConn = DbConn.createNative() if (
gConfig.connector_type == 'native') else DbConn.createRest()
try:
self._dbConn.open() # may throw taos.error.ProgrammingError: disconnected
self._dbConn.open() # may throw taos.error.ProgrammingError: disconnected
except taos.error.ProgrammingError as err:
# print("Error type: {}, msg: {}, value: {}".format(type(err), err.msg, err))
if ( err.msg == 'client disconnected' ): # cannot open DB connection
print("Cannot establish DB connection, please re-run script without parameter, and follow the instructions.")
if (err.msg == 'client disconnected'): # cannot open DB connection
print(
"Cannot establish DB connection, please re-run script without parameter, and follow the instructions.")
sys.exit(2)
else:
raise
except:
raise
except BaseException:
print("[=] Unexpected exception")
raise
raise
if resetDb:
self._dbConn.resetDb() # drop and recreate DB
if resetDb :
self._dbConn.resetDb() # drop and recreate DB
# Do this after dbConn is in proper shape
self._stateMachine = StateMechine(self._dbConn)
self._stateMachine = StateMechine(self._dbConn) # Do this after dbConn is in proper shape
def getDbConn(self):
return self._dbConn
def getStateMachine(self) -> StateMechine :
def getStateMachine(self) -> StateMechine:
return self._stateMachine
# def getState(self):
......@@ -1043,15 +1244,18 @@ class DbManager():
def setupLastTick(self):
t1 = datetime.datetime(2020, 6, 1)
t2 = datetime.datetime.now()
elSec = int(t2.timestamp() - t1.timestamp()) # maybe a very large number, takes 69 years to exceed Python int range
elSec2 = ( elSec % (8 * 12 * 30 * 24 * 60 * 60 / 500 ) ) * 500 # a number representing seconds within 10 years
# maybe a very large number, takes 69 years to exceed Python int range
elSec = int(t2.timestamp() - t1.timestamp())
elSec2 = (elSec % (8 * 12 * 30 * 24 * 60 * 60 / 500)) * \
500 # a number representing seconds within 10 years
# print("elSec = {}".format(elSec))
t3 = datetime.datetime(2012, 1, 1) # default "keep" is 10 years
t4 = datetime.datetime.fromtimestamp( t3.timestamp() + elSec2) # see explanation above
t3 = datetime.datetime(2012, 1, 1) # default "keep" is 10 years
t4 = datetime.datetime.fromtimestamp(
t3.timestamp() + elSec2) # see explanation above
logger.info("Setting up TICKS to start from: {}".format(t4))
return t4
def pickAndAllocateTable(self): # pick any table, and "use" it
def pickAndAllocateTable(self): # pick any table, and "use" it
return self.tableNumQueue.pickAndAllocate()
def addTable(self):
......@@ -1063,15 +1267,16 @@ class DbManager():
def getFixedSuperTableName(cls):
return "fs_table"
def releaseTable(self, i): # return the table back, so others can use it
def releaseTable(self, i): # return the table back, so others can use it
self.tableNumQueue.release(i)
def getNextTick(self):
with self._lock: # prevent duplicate tick
if Dice.throw(10) == 0 : # 1 in 10 chance
with self._lock: # prevent duplicate tick
if Dice.throw(10) == 0: # 1 in 10 chance
return self._lastTick + datetime.timedelta(0, -100)
else: # regular
self._lastTick += datetime.timedelta(0, 1) # add one second to it
else: # regular
# add one second to it
self._lastTick += datetime.timedelta(0, 1)
return self._lastTick
def getNextInt(self):
......@@ -1080,29 +1285,31 @@ class DbManager():
return self._lastInt
def getNextBinary(self):
return "Beijing_Shanghai_Los_Angeles_New_York_San_Francisco_Chicago_Beijing_Shanghai_Los_Angeles_New_York_San_Francisco_Chicago_{}".format(self.getNextInt())
return "Beijing_Shanghai_Los_Angeles_New_York_San_Francisco_Chicago_Beijing_Shanghai_Los_Angeles_New_York_San_Francisco_Chicago_{}".format(
self.getNextInt())
def getNextFloat(self):
return 0.9 + self.getNextInt()
def getTableNameToDelete(self):
tblNum = self.tableNumQueue.pop() # TODO: race condition!
if ( not tblNum ): # maybe false
tblNum = self.tableNumQueue.pop() # TODO: race condition!
if (not tblNum): # maybe false
return False
return "table_{}".format(tblNum)
def cleanUp(self):
self._dbConn.close()
self._dbConn.close()
class TaskExecutor():
class BoundedList:
def __init__(self, size = 10):
def __init__(self, size=10):
self._size = size
self._list = []
def add(self, n: int) :
if not self._list: # empty
def add(self, n: int):
if not self._list: # empty
self._list.append(n)
return
# now we should insert
......@@ -1110,22 +1317,22 @@ class TaskExecutor():
insPos = 0
for i in range(nItems):
insPos = i
if n <= self._list[i] : # smaller than this item, time to insert
break # found the insertion point
insPos += 1 # insert to the right
if n <= self._list[i]: # smaller than this item, time to insert
break # found the insertion point
insPos += 1 # insert to the right
if insPos == 0 : # except for the 1st item, # TODO: elimiate first item as gating item
return # do nothing
if insPos == 0: # except for the 1st item, # TODO: elimiate first item as gating item
return # do nothing
# print("Inserting at postion {}, value: {}".format(insPos, n))
self._list.insert(insPos, n) # insert
self._list.insert(insPos, n) # insert
newLen = len(self._list)
if newLen <= self._size :
return # do nothing
elif newLen == (self._size + 1) :
del self._list[0] # remove the first item
else :
if newLen <= self._size:
return # do nothing
elif newLen == (self._size + 1):
del self._list[0] # remove the first item
else:
raise RuntimeError("Corrupt Bounded List")
def __str__(self):
......@@ -1143,7 +1350,7 @@ class TaskExecutor():
def getCurStep(self):
return self._curStep
def execute(self, task: Task, wt: WorkerThread): # execute a task on a thread
def execute(self, task: Task, wt: WorkerThread): # execute a task on a thread
task.execute(wt)
def recordDataMark(self, n: int):
......@@ -1156,137 +1363,164 @@ class TaskExecutor():
# def logDebug(self, msg):
# logger.debug(" T[{}.x]: ".format(self._curStep) + msg)
class Task():
taskSn = 100
@classmethod
def allocTaskNum(cls):
Task.taskSn += 1 # IMPORTANT: cannot use cls.taskSn, since each sub class will have a copy
Task.taskSn += 1 # IMPORTANT: cannot use cls.taskSn, since each sub class will have a copy
# logger.debug("Allocating taskSN: {}".format(Task.taskSn))
return Task.taskSn
def __init__(self, dbManager: DbManager, execStats: ExecutionStats):
def __init__(self, dbManager: DbManager, execStats: ExecutionStats):
self._dbManager = dbManager
self._workerThread = None
self._workerThread = None
self._err = None
self._aborted = False
self._curStep = None
self._numRows = None # Number of rows affected
self._numRows = None # Number of rows affected
# Assign an incremental task serial number
# Assign an incremental task serial number
self._taskNum = self.allocTaskNum()
# logger.debug("Creating new task {}...".format(self._taskNum))
self._execStats = execStats
self._lastSql = "" # last SQL executed/attempted
self._lastSql = "" # last SQL executed/attempted
def isSuccess(self):
return self._err == None
return self._err is None
def isAborted(self):
return self._aborted
def clone(self): # TODO: why do we need this again?
def clone(self): # TODO: why do we need this again?
newTask = self.__class__(self._dbManager, self._execStats)
return newTask
def logDebug(self, msg):
self._workerThread.logDebug("Step[{}.{}] {}".format(self._curStep, self._taskNum, msg))
self._workerThread.logDebug(
"Step[{}.{}] {}".format(
self._curStep, self._taskNum, msg))
def logInfo(self, msg):
self._workerThread.logInfo("Step[{}.{}] {}".format(self._curStep, self._taskNum, msg))
self._workerThread.logInfo(
"Step[{}.{}] {}".format(
self._curStep, self._taskNum, msg))
def _executeInternal(self, te: TaskExecutor, wt: WorkerThread):
raise RuntimeError("To be implemeted by child classes, class name: {}".format(self.__class__.__name__))
raise RuntimeError(
"To be implemeted by child classes, class name: {}".format(
self.__class__.__name__))
def execute(self, wt: WorkerThread):
wt.verifyThreadSelf()
self._workerThread = wt # type: ignore
self._workerThread = wt # type: ignore
te = wt.getTaskExecutor()
self._curStep = te.getCurStep()
self.logDebug("[-] executing task {}...".format(self.__class__.__name__))
self.logDebug(
"[-] executing task {}...".format(self.__class__.__name__))
self._err = None
self._execStats.beginTaskType(self.__class__.__name__) # mark beginning
self._execStats.beginTaskType(
self.__class__.__name__) # mark beginning
try:
self._executeInternal(te, wt) # TODO: no return value?
self._executeInternal(te, wt) # TODO: no return value?
except taos.error.ProgrammingError as err:
errno2 = err.errno if (err.errno > 0) else 0x80000000 + err.errno # correct error scheme
if ( errno2 in [
0x05, # TSDB_CODE_RPC_NOT_READY
0x200, 0x360, 0x362, 0x36A, 0x36B, 0x36D, 0x381, 0x380, 0x383, 0x503,
0x510, # vnode not in ready state
errno2 = err.errno if (
err.errno > 0) else 0x80000000 + err.errno # correct error scheme
if (gConfig.continue_on_exception): # user choose to continue
self.logDebug(
"[=] Continue after TAOS exception: errno=0x{:X}, msg: {}, SQL: {}".format(
errno2, err, self._lastSql))
self._err = err
elif (errno2 in [
0x05, # TSDB_CODE_RPC_NOT_READY
0x200, 0x360, 0x362, 0x36A, 0x36B, 0x36D,
0x381, 0x380, 0x383,
0x386, # DB is being dropped?!
0x503,
0x510, # vnode not in ready state
0x600,
1000 # REST catch-all error
]) : # allowed errors
self.logDebug("[=] Acceptable Taos library exception: errno=0x{:X}, msg: {}, SQL: {}".format(errno2, err, self._lastSql))
1000 # REST catch-all error
]): # allowed errors
self.logDebug(
"[=] Acceptable Taos library exception: errno=0x{:X}, msg: {}, SQL: {}".format(
errno2, err, self._lastSql))
print("_", end="", flush=True)
self._err = err
self._err = err
else:
errMsg = "[=] Unexpected Taos library exception: errno=0x{:X}, msg: {}, SQL: {}".format(errno2, err, self._lastSql)
errMsg = "[=] Unexpected Taos library exception: errno=0x{:X}, msg: {}, SQL: {}".format(
errno2, err, self._lastSql)
self.logDebug(errMsg)
if gConfig.debug :
if gConfig.debug:
# raise # so that we see full stack
traceback.print_exc()
print("\n\n----------------------------\nProgram ABORTED Due to Unexpected TAOS Error: \n\n{}\n".format(errMsg) +
print(
"\n\n----------------------------\nProgram ABORTED Due to Unexpected TAOS Error: \n\n{}\n".format(errMsg) +
"----------------------------\n")
# sys.exit(-1)
self._err = err
self._aborted = True
except Exception as e :
except Exception as e:
self.logInfo("Non-TAOS exception encountered")
self._err = e
self._err = e
self._aborted = True
traceback.print_exc()
except BaseException as e :
except BaseException as e:
self.logInfo("Python base exception encountered")
self._err = e
self._err = e
self._aborted = True
traceback.print_exc()
except :
self.logDebug("[=] Unexpected exception, SQL: {}".format(self._lastSql))
except BaseException:
self.logDebug(
"[=] Unexpected exception, SQL: {}".format(
self._lastSql))
raise
self._execStats.endTaskType(self.__class__.__name__, self.isSuccess())
self.logDebug("[X] task execution completed, {}, status: {}".format(self.__class__.__name__, "Success" if self.isSuccess() else "Failure"))
self._execStats.incExecCount(self.__class__.__name__, self.isSuccess()) # TODO: merge with above.
self.logDebug("[X] task execution completed, {}, status: {}".format(
self.__class__.__name__, "Success" if self.isSuccess() else "Failure"))
# TODO: merge with above.
self._execStats.incExecCount(self.__class__.__name__, self.isSuccess())
def execSql(self, sql):
self._lastSql = sql
return self._dbManager.execute(sql)
def execWtSql(self, wt: WorkerThread, sql): # execute an SQL on the worker thread
def execWtSql(self, wt: WorkerThread, sql): # execute an SQL on the worker thread
self._lastSql = sql
return wt.execSql(sql)
def queryWtSql(self, wt: WorkerThread, sql): # execute an SQL on the worker thread
def queryWtSql(self, wt: WorkerThread, sql): # execute an SQL on the worker thread
self._lastSql = sql
return wt.querySql(sql)
def getQueryResult(self, wt: WorkerThread): # execute an SQL on the worker thread
def getQueryResult(self, wt: WorkerThread): # execute an SQL on the worker thread
return wt.getQueryResult()
class ExecutionStats:
def __init__(self):
self._execTimes: Dict[str, [int, int]] = {} # total/success times for a task
# total/success times for a task
self._execTimes: Dict[str, [int, int]] = {}
self._tasksInProgress = 0
self._lock = threading.Lock()
self._firstTaskStartTime = None
self._execStartTime = None
self._elapsedTime = 0.0 # total elapsed time
self._accRunTime = 0.0 # accumulated run time
self._elapsedTime = 0.0 # total elapsed time
self._accRunTime = 0.0 # accumulated run time
self._failed = False
self._failureReason = None
def __str__(self):
return "[ExecStats: _failed={}, _failureReason={}".format(self._failed, self._failureReason)
return "[ExecStats: _failed={}, _failureReason={}".format(
self._failed, self._failureReason)
def isFailed(self):
return self._failed == True
return self._failed
def startExec(self):
self._execStartTime = time.time()
......@@ -1294,24 +1528,24 @@ class ExecutionStats:
def endExec(self):
self._elapsedTime = time.time() - self._execStartTime
def incExecCount(self, klassName, isSuccess): # TODO: add a lock here
def incExecCount(self, klassName, isSuccess): # TODO: add a lock here
if klassName not in self._execTimes:
self._execTimes[klassName] = [0, 0]
t = self._execTimes[klassName] # tuple for the data
t[0] += 1 # index 0 has the "total" execution times
t = self._execTimes[klassName] # tuple for the data
t[0] += 1 # index 0 has the "total" execution times
if isSuccess:
t[1] += 1 # index 1 has the "success" execution times
t[1] += 1 # index 1 has the "success" execution times
def beginTaskType(self, klassName):
with self._lock:
if self._tasksInProgress == 0 : # starting a new round
self._firstTaskStartTime = time.time() # I am now the first task
if self._tasksInProgress == 0: # starting a new round
self._firstTaskStartTime = time.time() # I am now the first task
self._tasksInProgress += 1
def endTaskType(self, klassName, isSuccess):
with self._lock:
self._tasksInProgress -= 1
if self._tasksInProgress == 0 : # all tasks have stopped
if self._tasksInProgress == 0: # all tasks have stopped
self._accRunTime += (time.time() - self._firstTaskStartTime)
self._firstTaskStartTime = None
......@@ -1320,23 +1554,36 @@ class ExecutionStats:
self._failureReason = reason
def printStats(self):
logger.info("----------------------------------------------------------------------")
logger.info("| Crash_Gen test {}, with the following stats:".
format("FAILED (reason: {})".format(self._failureReason) if self._failed else "SUCCEEDED"))
logger.info(
"----------------------------------------------------------------------")
logger.info(
"| Crash_Gen test {}, with the following stats:". format(
"FAILED (reason: {})".format(
self._failureReason) if self._failed else "SUCCEEDED"))
logger.info("| Task Execution Times (success/total):")
execTimesAny = 0
for k, n in self._execTimes.items():
for k, n in self._execTimes.items():
execTimesAny += n[0]
logger.info("| {0:<24}: {1}/{2}".format(k,n[1],n[0]))
logger.info("| Total Tasks Executed (success or not): {} ".format(execTimesAny))
logger.info("| Total Tasks In Progress at End: {}".format(self._tasksInProgress))
logger.info("| Total Task Busy Time (elapsed time when any task is in progress): {:.3f} seconds".format(self._accRunTime))
logger.info("| Average Per-Task Execution Time: {:.3f} seconds".format(self._accRunTime/execTimesAny))
logger.info("| Total Elapsed Time (from wall clock): {:.3f} seconds".format(self._elapsedTime))
logger.info("| Top numbers written: {}".format(TaskExecutor.getBoundedList()))
logger.info("----------------------------------------------------------------------")
logger.info("| {0:<24}: {1}/{2}".format(k, n[1], n[0]))
logger.info(
"| Total Tasks Executed (success or not): {} ".format(execTimesAny))
logger.info(
"| Total Tasks In Progress at End: {}".format(
self._tasksInProgress))
logger.info(
"| Total Task Busy Time (elapsed time when any task is in progress): {:.3f} seconds".format(
self._accRunTime))
logger.info(
"| Average Per-Task Execution Time: {:.3f} seconds".format(self._accRunTime / execTimesAny))
logger.info(
"| Total Elapsed Time (from wall clock): {:.3f} seconds".format(
self._elapsedTime))
logger.info(
"| Top numbers written: {}".format(
TaskExecutor.getBoundedList()))
logger.info(
"----------------------------------------------------------------------")
class StateTransitionTask(Task):
......@@ -1346,12 +1593,12 @@ class StateTransitionTask(Task):
SMALL_NUMBER_OF_RECORDS = 3
@classmethod
def getInfo(cls): # each sub class should supply their own information
def getInfo(cls): # each sub class should supply their own information
raise RuntimeError("Overriding method expected")
_endState = None
_endState = None
@classmethod
def getEndState(cls): # TODO: optimize by calling it fewer times
def getEndState(cls): # TODO: optimize by calling it fewer times
raise RuntimeError("Overriding method expected")
# @classmethod
......@@ -1373,18 +1620,20 @@ class StateTransitionTask(Task):
def execute(self, wt: WorkerThread):
super().execute(wt)
class TaskCreateDb(StateTransitionTask):
@classmethod
def getEndState(cls):
return StateDbOnly()
return StateDbOnly()
@classmethod
def canBeginFrom(cls, state: AnyState):
return state.canCreateDb()
def _executeInternal(self, te: TaskExecutor, wt: WorkerThread):
self.execWtSql(wt, "create database db")
self.execWtSql(wt, "create database db")
class TaskDropDb(StateTransitionTask):
@classmethod
......@@ -1399,6 +1648,7 @@ class TaskDropDb(StateTransitionTask):
self.execWtSql(wt, "drop database db")
logger.debug("[OPS] database dropped at {}".format(time.time()))
class TaskCreateSuperTable(StateTransitionTask):
@classmethod
def getEndState(cls):
......@@ -1409,115 +1659,135 @@ class TaskCreateSuperTable(StateTransitionTask):
return state.canCreateFixedSuperTable()
def _executeInternal(self, te: TaskExecutor, wt: WorkerThread):
if not wt.dbInUse(): # no DB yet, to the best of our knowledge
if not wt.dbInUse(): # no DB yet, to the best of our knowledge
logger.debug("Skipping task, no DB yet")
return
tblName = self._dbManager.getFixedSuperTableName()
tblName = self._dbManager.getFixedSuperTableName()
# wt.execSql("use db") # should always be in place
self.execWtSql(wt, "create table db.{} (ts timestamp, speed int) tags (b binary(200), f float) ".format(tblName))
# No need to create the regular tables, INSERT will do that automatically
self.execWtSql(
wt,
"create table db.{} (ts timestamp, speed int) tags (b binary(200), f float) ".format(tblName))
# No need to create the regular tables, INSERT will do that
# automatically
class TaskReadData(StateTransitionTask):
@classmethod
def getEndState(cls):
return None # meaning doesn't affect state
return None # meaning doesn't affect state
@classmethod
def canBeginFrom(cls, state: AnyState):
return state.canReadData()
def _executeInternal(self, te: TaskExecutor, wt: WorkerThread):
sTbName = self._dbManager.getFixedSuperTableName()
self.queryWtSql(wt, "select TBNAME from db.{}".format(sTbName)) # TODO: analyze result set later
sTbName = self._dbManager.getFixedSuperTableName()
self.queryWtSql(wt, "select TBNAME from db.{}".format(
sTbName)) # TODO: analyze result set later
if random.randrange(5) == 0 : # 1 in 5 chance, simulate a broken connection. TODO: break connection in all situations
if random.randrange(
5) == 0: # 1 in 5 chance, simulate a broken connection. TODO: break connection in all situations
wt.getDbConn().close()
wt.getDbConn().open()
else:
rTables = self.getQueryResult(wt) # wt.getDbConn().getQueryResult()
# wt.getDbConn().getQueryResult()
rTables = self.getQueryResult(wt)
# print("rTables[0] = {}, type = {}".format(rTables[0], type(rTables[0])))
for rTbName in rTables : # regular tables
for rTbName in rTables: # regular tables
self.execWtSql(wt, "select * from db.{}".format(rTbName[0]))
# tdSql.query(" cars where tbname in ('carzero', 'carone')")
class TaskDropSuperTable(StateTransitionTask):
@classmethod
def getEndState(cls):
return StateDbOnly()
return StateDbOnly()
@classmethod
def canBeginFrom(cls, state: AnyState):
return state.canDropFixedSuperTable()
def _executeInternal(self, te: TaskExecutor, wt: WorkerThread):
# 1/2 chance, we'll drop the regular tables one by one, in a randomized sequence
if Dice.throw(2) == 0 :
tblSeq = list(range(2 + (self.LARGE_NUMBER_OF_TABLES if gConfig.larger_data else self.SMALL_NUMBER_OF_TABLES)))
random.shuffle(tblSeq)
tickOutput = False # if we have spitted out a "d" character for "drop regular table"
# 1/2 chance, we'll drop the regular tables one by one, in a randomized
# sequence
if Dice.throw(2) == 0:
tblSeq = list(range(
2 + (self.LARGE_NUMBER_OF_TABLES if gConfig.larger_data else self.SMALL_NUMBER_OF_TABLES)))
random.shuffle(tblSeq)
tickOutput = False # if we have spitted out a "d" character for "drop regular table"
isSuccess = True
for i in tblSeq:
regTableName = self.getRegTableName(i); # "db.reg_table_{}".format(i)
for i in tblSeq:
regTableName = self.getRegTableName(
i) # "db.reg_table_{}".format(i)
try:
self.execWtSql(wt, "drop table {}".format(regTableName)) # nRows always 0, like MySQL
except taos.error.ProgrammingError as err:
errno2 = err.errno if (err.errno > 0) else 0x80000000 + err.errno # correcting for strange error number scheme
if ( errno2 in [0x362]) : # mnode invalid table name
self.execWtSql(wt, "drop table {}".format(
regTableName)) # nRows always 0, like MySQL
except taos.error.ProgrammingError as err:
# correcting for strange error number scheme
errno2 = err.errno if (
err.errno > 0) else 0x80000000 + err.errno
if (errno2 in [0x362]): # mnode invalid table name
isSuccess = False
logger.debug("[DB] Acceptable error when dropping a table")
continue # try to delete next regular table
logger.debug(
"[DB] Acceptable error when dropping a table")
continue # try to delete next regular table
if (not tickOutput):
tickOutput = True # Print only one time
if isSuccess :
tickOutput = True # Print only one time
if isSuccess:
print("d", end="", flush=True)
else:
print("f", end="", flush=True)
print("f", end="", flush=True)
# Drop the super table itself
tblName = self._dbManager.getFixedSuperTableName()
tblName = self._dbManager.getFixedSuperTableName()
self.execWtSql(wt, "drop table db.{}".format(tblName))
class TaskAlterTags(StateTransitionTask):
@classmethod
def getEndState(cls):
return None # meaning doesn't affect state
return None # meaning doesn't affect state
@classmethod
def canBeginFrom(cls, state: AnyState):
return state.canDropFixedSuperTable() # if we can drop it, we can alter tags
return state.canDropFixedSuperTable() # if we can drop it, we can alter tags
def _executeInternal(self, te: TaskExecutor, wt: WorkerThread):
tblName = self._dbManager.getFixedSuperTableName()
tblName = self._dbManager.getFixedSuperTableName()
dice = Dice.throw(4)
if dice == 0 :
if dice == 0:
sql = "alter table db.{} add tag extraTag int".format(tblName)
elif dice == 1 :
elif dice == 1:
sql = "alter table db.{} drop tag extraTag".format(tblName)
elif dice == 2 :
elif dice == 2:
sql = "alter table db.{} drop tag newTag".format(tblName)
else: # dice == 3
sql = "alter table db.{} change tag extraTag newTag".format(tblName)
else: # dice == 3
sql = "alter table db.{} change tag extraTag newTag".format(
tblName)
self.execWtSql(wt, sql)
class TaskAddData(StateTransitionTask):
activeTable : Set[int] = set() # Track which table is being actively worked on
# Track which table is being actively worked on
activeTable: Set[int] = set()
# We use these two files to record operations to DB, useful for power-off tests
# We use these two files to record operations to DB, useful for power-off
# tests
fAddLogReady = None
fAddLogDone = None
@classmethod
def prepToRecordOps(cls):
if gConfig.record_ops :
if ( cls.fAddLogReady == None ):
logger.info("Recording in a file operations to be performed...")
if gConfig.record_ops:
if (cls.fAddLogReady is None):
logger.info(
"Recording in a file operations to be performed...")
cls.fAddLogReady = open("add_log_ready.txt", "w")
if ( cls.fAddLogDone == None ):
if (cls.fAddLogDone is None):
logger.info("Recording in a file operations completed...")
cls.fAddLogDone = open("add_log_done.txt", "w")
......@@ -1528,78 +1798,92 @@ class TaskAddData(StateTransitionTask):
@classmethod
def canBeginFrom(cls, state: AnyState):
return state.canAddData()
def _executeInternal(self, te: TaskExecutor, wt: WorkerThread):
ds = self._dbManager
# wt.execSql("use db") # TODO: seems to be an INSERT bug to require this
tblSeq = list(range(self.LARGE_NUMBER_OF_TABLES if gConfig.larger_data else self.SMALL_NUMBER_OF_TABLES))
random.shuffle(tblSeq)
for i in tblSeq:
if ( i in self.activeTable ): # wow already active
# logger.info("Concurrent data insertion into table: {}".format(i))
# print("ct({})".format(i), end="", flush=True) # Concurrent insertion into table
# wt.execSql("use db") # TODO: seems to be an INSERT bug to require
# this
tblSeq = list(
range(
self.LARGE_NUMBER_OF_TABLES if gConfig.larger_data else self.SMALL_NUMBER_OF_TABLES))
random.shuffle(tblSeq)
for i in tblSeq:
if (i in self.activeTable): # wow already active
# logger.info("Concurrent data insertion into table: {}".format(i))
# print("ct({})".format(i), end="", flush=True) # Concurrent
# insertion into table
print("x", end="", flush=True)
else:
self.activeTable.add(i) # marking it active
# No need to shuffle data sequence, unless later we decide to do non-increment insertion
regTableName = self.getRegTableName(i); # "db.reg_table_{}".format(i)
for j in range(self.LARGE_NUMBER_OF_RECORDS if gConfig.larger_data else self.SMALL_NUMBER_OF_RECORDS) : # number of records per table
nextInt = ds.getNextInt()
self.activeTable.add(i) # marking it active
# No need to shuffle data sequence, unless later we decide to do
# non-increment insertion
regTableName = self.getRegTableName(
i) # "db.reg_table_{}".format(i)
for j in range(
self.LARGE_NUMBER_OF_RECORDS if gConfig.larger_data else self.SMALL_NUMBER_OF_RECORDS): # number of records per table
nextInt = ds.getNextInt()
if gConfig.record_ops:
self.prepToRecordOps()
self.fAddLogReady.write("Ready to write {} to {}\n".format(nextInt, regTableName))
self.fAddLogReady.write(
"Ready to write {} to {}\n".format(
nextInt, regTableName))
self.fAddLogReady.flush()
os.fsync(self.fAddLogReady)
sql = "insert into {} using {} tags ('{}', {}) values ('{}', {});".format(
regTableName,
ds.getFixedSuperTableName(),
regTableName,
ds.getFixedSuperTableName(),
ds.getNextBinary(), ds.getNextFloat(),
ds.getNextTick(), nextInt)
self.execWtSql(wt, sql)
# Successfully wrote the data into the DB, let's record it somehow
self.execWtSql(wt, sql)
# Successfully wrote the data into the DB, let's record it
# somehow
te.recordDataMark(nextInt)
if gConfig.record_ops:
self.fAddLogDone.write("Wrote {} to {}\n".format(nextInt, regTableName))
self.fAddLogDone.write(
"Wrote {} to {}\n".format(
nextInt, regTableName))
self.fAddLogDone.flush()
os.fsync(self.fAddLogDone)
self.activeTable.discard(i) # not raising an error, unlike remove
self.activeTable.discard(i) # not raising an error, unlike remove
# Deterministic random number generator
class Dice():
seeded = False # static, uninitialized
seeded = False # static, uninitialized
@classmethod
def seed(cls, s): # static
def seed(cls, s): # static
if (cls.seeded):
raise RuntimeError("Cannot seed the random generator more than once")
raise RuntimeError(
"Cannot seed the random generator more than once")
cls.verifyRNG()
random.seed(s)
cls.seeded = True # TODO: protect against multi-threading
@classmethod
def verifyRNG(cls): # Verify that the RNG is determinstic
def verifyRNG(cls): # Verify that the RNG is determinstic
random.seed(0)
x1 = random.randrange(0, 1000)
x2 = random.randrange(0, 1000)
x3 = random.randrange(0, 1000)
if ( x1 != 864 or x2!=394 or x3!=776 ):
if (x1 != 864 or x2 != 394 or x3 != 776):
raise RuntimeError("System RNG is not deterministic")
@classmethod
def throw(cls, stop): # get 0 to stop-1
def throw(cls, stop): # get 0 to stop-1
return cls.throwRange(0, stop)
@classmethod
def throwRange(cls, start, stop): # up to stop-1
if ( not cls.seeded ):
def throwRange(cls, start, stop): # up to stop-1
if (not cls.seeded):
raise RuntimeError("Cannot throw dice before seeding it")
return random.randrange(start, stop)
class LoggingFilter(logging.Filter):
def filter(self, record: logging.LogRecord):
if ( record.levelno >= logging.INFO ) :
return True # info or above always log
if (record.levelno >= logging.INFO):
return True # info or above always log
# Commenting out below to adjust...
......@@ -1607,20 +1891,23 @@ class LoggingFilter(logging.Filter):
# return False
return True
class MyLoggingAdapter(logging.LoggerAdapter):
class MyLoggingAdapter(logging.LoggerAdapter):
def process(self, msg, kwargs):
return "[{}]{}".format(threading.get_ident() % 10000, msg), kwargs
# return '[%s] %s' % (self.extra['connid'], msg), kwargs
class SvcManager:
class SvcManager:
def __init__(self):
print("Starting TDengine Service Manager")
signal.signal(signal.SIGTERM, self.sigIntHandler)
signal.signal(signal.SIGINT, self.sigIntHandler)
signal.signal(signal.SIGUSR1, self.sigUsrHandler) # different handler!
signal.signal(signal.SIGINT, self.sigIntHandler)
signal.signal(signal.SIGUSR1, self.sigUsrHandler) # different handler!
self.inSigHandler = False
# self._status = MainExec.STATUS_RUNNING # set inside _startTaosService()
# self._status = MainExec.STATUS_RUNNING # set inside
# _startTaosService()
self.svcMgrThread = None
def _doMenu(self):
......@@ -1631,30 +1918,32 @@ class SvcManager:
print("2: Terminate")
print("3: Restart")
# Remember to update the if range below
# print("Enter Choice: ", end="", flush=True)
# print("Enter Choice: ", end="", flush=True)
while choice == "":
choice = input("Enter Choice: ")
if choice != "":
break # done with reading repeated input
if choice in ["1", "2", "3"]:
break # we are done with whole method
break # done with reading repeated input
if choice in ["1", "2", "3"]:
break # we are done with whole method
print("Invalid choice, please try again.")
choice = "" # reset
choice = "" # reset
return choice
def sigUsrHandler(self, signalNumber, frame) :
def sigUsrHandler(self, signalNumber, frame):
print("Interrupting main thread execution upon SIGUSR1")
if self.inSigHandler : # already
if self.inSigHandler: # already
print("Ignoring repeated SIG...")
return # do nothing if it's already not running
return # do nothing if it's already not running
self.inSigHandler = True
choice = self._doMenu()
if choice == "1" :
self.sigHandlerResume() # TODO: can the sub-process be blocked due to us not reading from queue?
elif choice == "2" :
if choice == "1":
# TODO: can the sub-process be blocked due to us not reading from
# queue?
self.sigHandlerResume()
elif choice == "2":
self.stopTaosService()
elif choice == "3" :
elif choice == "3":
self.stopTaosService()
self.startTaosService()
else:
......@@ -1664,48 +1953,52 @@ class SvcManager:
def sigIntHandler(self, signalNumber, frame):
print("Sig INT Handler starting...")
if self.inSigHandler :
if self.inSigHandler:
print("Ignoring repeated SIG_INT...")
return
self.inSigHandler = True
self.stopTaosService()
print("INT signal handler returning...")
self.stopTaosService()
print("INT signal handler returning...")
self.inSigHandler = False
def sigHandlerResume(self) :
def sigHandlerResume(self):
print("Resuming TDengine service manager thread (main thread)...\n\n")
def _checkServiceManagerThread(self):
if self.svcMgrThread: # valid svc mgr thread
if self.svcMgrThread.isStopped(): # done?
self.svcMgrThread.procIpcBatch() # one last time. TODO: appropriate?
self.svcMgrThread = None # no more
if self.svcMgrThread: # valid svc mgr thread
if self.svcMgrThread.isStopped(): # done?
self.svcMgrThread.procIpcBatch() # one last time. TODO: appropriate?
self.svcMgrThread = None # no more
def _procIpcAll(self):
while self.svcMgrThread : # for as long as the svc mgr thread is still here
self.svcMgrThread.procIpcBatch() # regular processing,
time.sleep(0.5) # pause, before next round
self._checkServiceManagerThread()
print("Service Manager Thread (with subprocess) has ended, main thread now exiting...")
def startTaosService(self):
while self.svcMgrThread: # for as long as the svc mgr thread is still here
self.svcMgrThread.procIpcBatch() # regular processing,
time.sleep(0.5) # pause, before next round
self._checkServiceManagerThread()
print(
"Service Manager Thread (with subprocess) has ended, main thread now exiting...")
def startTaosService(self):
if self.svcMgrThread:
raise RuntimeError("Cannot start TAOS service when one may already be running")
self.svcMgrThread = ServiceManagerThread() # create the object
raise RuntimeError(
"Cannot start TAOS service when one may already be running")
self.svcMgrThread = ServiceManagerThread() # create the object
self.svcMgrThread.start()
print("TAOS service started, printing out output...")
self.svcMgrThread.procIpcBatch(trimToTarget=10, forceOutput=True) # for printing 10 lines
self.svcMgrThread.procIpcBatch(
trimToTarget=10,
forceOutput=True) # for printing 10 lines
print("TAOS service started")
def stopTaosService(self, outputLines = 20):
def stopTaosService(self, outputLines=20):
print("Terminating Service Manager Thread (SMT) execution...")
if not self.svcMgrThread:
raise RuntimeError("Unexpected empty svc mgr thread")
self.svcMgrThread.stop()
if self.svcMgrThread.isStopped():
self.svcMgrThread.procIpcBatch(outputLines) # one last time
self.svcMgrThread = None
self.svcMgrThread.procIpcBatch(outputLines) # one last time
self.svcMgrThread = None
print("----- End of TDengine Service Output -----\n")
print("SMT execution terminated")
else:
......@@ -1713,16 +2006,17 @@ class SvcManager:
def run(self):
self.startTaosService()
self._procIpcAll() # pump/process all the messages
if self.svcMgrThread: # if sig handler hasn't destroyed it by now
self.stopTaosService() # should have started already
self._procIpcAll() # pump/process all the messages
if self.svcMgrThread: # if sig handler hasn't destroyed it by now
self.stopTaosService() # should have started already
class ServiceManagerThread:
MAX_QUEUE_SIZE = 10000
def __init__(self):
self._tdeSubProcess = None
self._thread = None
self._thread = None
self._status = None
def getStatus(self):
......@@ -1740,98 +2034,107 @@ class ServiceManagerThread:
# Start the thread (with sub process), and wait for the sub service
# to become fully operational
def start(self):
if self._thread :
def start(self):
if self._thread:
raise RuntimeError("Unexpected _thread")
if self._tdeSubProcess :
if self._tdeSubProcess:
raise RuntimeError("TDengine sub process already created/running")
self._status = MainExec.STATUS_STARTING
self._tdeSubProcess = TdeSubProcess()
self._tdeSubProcess = TdeSubProcess()
self._tdeSubProcess.start()
self._ipcQueue = Queue()
self._thread = threading.Thread(
target=self.svcOutputReader,
target=self.svcOutputReader,
args=(self._tdeSubProcess.getStdOut(), self._ipcQueue))
self._thread.daemon = True # thread dies with the program
self._thread.daemon = True # thread dies with the program
self._thread.start()
# wait for service to start
for i in range(0, 10) :
for i in range(0, 10):
time.sleep(1.0)
# self.procIpcBatch() # don't pump message during start up
print("_zz_", end="", flush=True)
if self._status == MainExec.STATUS_RUNNING :
if self._status == MainExec.STATUS_RUNNING:
logger.info("[] TDengine service READY to process requests")
return # now we've started
raise RuntimeError("TDengine service did not start successfully") # TODO: handle this better?
return # now we've started
# TODO: handle this better?
raise RuntimeError("TDengine service did not start successfully")
def stop(self):
# can be called from both main thread or signal handler
print("Terminating TDengine service running as the sub process...")
if self.isStopped():
print("Service already stopped")
return
return
if self.isStopping():
print("Service is already being stopped")
return
# Linux will send Control-C generated SIGINT to the TDengine process already, ref: https://unix.stackexchange.com/questions/176235/fork-and-how-signals-are-delivered-to-processes
if not self._tdeSubProcess :
# Linux will send Control-C generated SIGINT to the TDengine process
# already, ref:
# https://unix.stackexchange.com/questions/176235/fork-and-how-signals-are-delivered-to-processes
if not self._tdeSubProcess:
raise RuntimeError("sub process object missing")
self._status = MainExec.STATUS_STOPPING
self._tdeSubProcess.stop()
if self._tdeSubProcess.isRunning(): # still running
print("FAILED to stop sub process, it is still running... pid = {}".format(self.subProcess.pid))
if self._tdeSubProcess.isRunning(): # still running
print(
"FAILED to stop sub process, it is still running... pid = {}".format(
self.subProcess.pid))
else:
self._tdeSubProcess = None # not running any more
self.join() # stop the thread, change the status, etc.
self._tdeSubProcess = None # not running any more
self.join() # stop the thread, change the status, etc.
def join(self):
# TODO: sanity check
if not self.isStopping():
raise RuntimeError("Unexpected status when ending svc mgr thread: {}".format(self._status))
raise RuntimeError(
"Unexpected status when ending svc mgr thread: {}".format(
self._status))
if self._thread :
if self._thread:
self._thread.join()
self._thread = None
self._thread = None
self._status = MainExec.STATUS_STOPPED
else :
else:
print("Joining empty thread, doing nothing")
def _trimQueue(self, targetSize):
if targetSize <= 0:
return # do nothing
return # do nothing
q = self._ipcQueue
if (q.qsize() <= targetSize ) : # no need to trim
if (q.qsize() <= targetSize): # no need to trim
return
logger.debug("Triming IPC queue to target size: {}".format(targetSize))
itemsToTrim = q.qsize() - targetSize
for i in range(0, itemsToTrim) :
for i in range(0, itemsToTrim):
try:
q.get_nowait()
except Empty:
break # break out of for loop, no more trimming
break # break out of for loop, no more trimming
TD_READY_MSG = "TDengine is initialized successfully"
def procIpcBatch(self, trimToTarget = 0, forceOutput = False):
self._trimQueue(trimToTarget) # trim if necessary
# Process all the output generated by the underlying sub process, managed by IO thread
def procIpcBatch(self, trimToTarget=0, forceOutput=False):
self._trimQueue(trimToTarget) # trim if necessary
# Process all the output generated by the underlying sub process,
# managed by IO thread
print("<", end="", flush=True)
while True :
try:
line = self._ipcQueue.get_nowait() # getting output at fast speed
while True:
try:
line = self._ipcQueue.get_nowait() # getting output at fast speed
self._printProgress("_o")
except Empty:
# time.sleep(2.3) # wait only if there's no output
# no more output
print(".>", end="", flush=True)
return # we are done with THIS BATCH
else: # got line, printing out
return # we are done with THIS BATCH
else: # got line, printing out
if forceOutput:
logger.info(line)
else:
......@@ -1839,8 +2142,9 @@ class ServiceManagerThread:
print(">", end="", flush=True)
_ProgressBars = ["--", "//", "||", "\\\\"]
def _printProgress(self, msg): # TODO: assuming 2 chars
print(msg, end="", flush=True)
def _printProgress(self, msg): # TODO: assuming 2 chars
print(msg, end="", flush=True)
pBar = self._ProgressBars[Dice.throw(4)]
print(pBar, end="", flush=True)
print('\b\b\b\b', end="", flush=True)
......@@ -1848,29 +2152,33 @@ class ServiceManagerThread:
def svcOutputReader(self, out: IO, queue):
# Important Reference: https://stackoverflow.com/questions/375427/non-blocking-read-on-a-subprocess-pipe-in-python
# print("This is the svcOutput Reader...")
# for line in out :
# for line in out :
for line in iter(out.readline, b''):
# print("Finished reading a line: {}".format(line))
# print("Adding item to queue...")
line = line.decode("utf-8").rstrip()
queue.put(line) # This might block, and then causing "out" buffer to block
# This might block, and then causing "out" buffer to block
queue.put(line)
self._printProgress("_i")
if self._status == MainExec.STATUS_STARTING : # we are starting, let's see if we have started
if line.find(self.TD_READY_MSG) != -1 : # found
self._status = MainExec.STATUS_RUNNING
if self._status == MainExec.STATUS_STARTING: # we are starting, let's see if we have started
if line.find(self.TD_READY_MSG) != -1: # found
self._status = MainExec.STATUS_RUNNING
# Trim the queue if necessary: TODO: try this 1 out of 10 times
self._trimQueue(self.MAX_QUEUE_SIZE * 9 // 10) # trim to 90% size
self._trimQueue(self.MAX_QUEUE_SIZE * 9 // 10) # trim to 90% size
if self.isStopping() : # TODO: use thread status instead
print("_w", end="", flush=True) # WAITING for stopping sub process to finish its outptu
if self.isStopping(): # TODO: use thread status instead
# WAITING for stopping sub process to finish its outptu
print("_w", end="", flush=True)
# queue.put(line)
print("\nNo more output from IO thread managing TDengine service") # meaning sub process must have died
# meaning sub process must have died
print("\nNo more output from IO thread managing TDengine service")
out.close()
class TdeSubProcess:
class TdeSubProcess:
def __init__(self):
self.subProcess = None
......@@ -1878,20 +2186,39 @@ class TdeSubProcess:
return self.subProcess.stdout
def isRunning(self):
return self.subProcess != None
return self.subProcess is not None
def getBuildPath(self):
selfPath = os.path.dirname(os.path.realpath(__file__))
if ("community" in selfPath):
projPath = selfPath[:selfPath.find("communit")]
else:
projPath = selfPath[:selfPath.find("tests")]
for root, dirs, files in os.walk(projPath):
if ("taosd" in files):
rootRealPath = os.path.dirname(os.path.realpath(root))
if ("packaging" not in rootRealPath):
buildPath = root[:len(root) - len("/build/bin")]
break
return buildPath
def start(self):
ON_POSIX = 'posix' in sys.builtin_module_names
svcCmd = ['../../build/build/bin/taosd', '-c', '../../build/test/cfg']
taosdPath = self.getBuildPath() + "/build/bin/taosd"
cfgPath = self.getBuildPath() + "/test/cfg"
svcCmd = [taosdPath, '-c', cfgPath]
# svcCmd = ['vmstat', '1']
if self.subProcess : # already there
if self.subProcess: # already there
raise RuntimeError("Corrupt process state")
self.subProcess = subprocess.Popen(
svcCmd,
stdout=subprocess.PIPE,
svcCmd,
stdout=subprocess.PIPE,
# bufsize=1, # not supported in binary mode
close_fds=ON_POSIX) # had text=True, which interferred with reading EOF
close_fds=ON_POSIX) # had text=True, which interferred with reading EOF
def stop(self):
if not self.subProcess:
......@@ -1899,12 +2226,15 @@ class TdeSubProcess:
return
retCode = self.subProcess.poll()
if retCode : # valid return code, process ended
if retCode: # valid return code, process ended
self.subProcess = None
else: # process still alive, let's interrupt it
print("Sub process is running, sending SIG_INT and waiting for it to terminate...")
self.subProcess.send_signal(signal.SIGINT) # sub process should end, then IPC queue should end, causing IO thread to end
try :
else: # process still alive, let's interrupt it
print(
"Sub process is running, sending SIG_INT and waiting for it to terminate...")
# sub process should end, then IPC queue should end, causing IO
# thread to end
self.subProcess.send_signal(signal.SIGINT)
try:
self.subProcess.wait(10)
except subprocess.TimeoutExpired as err:
print("Time out waiting for TDengine service process to exit")
......@@ -1912,6 +2242,7 @@ class TdeSubProcess:
print("TDengine service process terminated successfully from SIG_INT")
self.subProcess = None
class ClientManager:
def __init__(self):
print("Starting service manager")
......@@ -1922,76 +2253,78 @@ class ClientManager:
self.tc = None
def sigIntHandler(self, signalNumber, frame):
if self._status != MainExec.STATUS_RUNNING :
if self._status != MainExec.STATUS_RUNNING:
print("Ignoring repeated SIGINT...")
return # do nothing if it's already not running
self._status = MainExec.STATUS_STOPPING # immediately set our status
return # do nothing if it's already not running
self._status = MainExec.STATUS_STOPPING # immediately set our status
print("Terminating program...")
self.tc.requestToStop()
def _printLastNumbers(self): # to verify data durability
def _printLastNumbers(self): # to verify data durability
dbManager = DbManager(resetDb=False)
dbc = dbManager.getDbConn()
if dbc.query("show databases") == 0 : # no databae
if dbc.query("show databases") == 0: # no databae
return
if dbc.query("show tables") == 0 : # no tables
if dbc.query("show tables") == 0: # no tables
return
dbc.execute("use db")
sTbName = dbManager.getFixedSuperTableName()
sTbName = dbManager.getFixedSuperTableName()
# get all regular tables
dbc.query("select TBNAME from db.{}".format(sTbName)) # TODO: analyze result set later
# TODO: analyze result set later
dbc.query("select TBNAME from db.{}".format(sTbName))
rTables = dbc.getQueryResult()
bList = TaskExecutor.BoundedList()
for rTbName in rTables : # regular tables
for rTbName in rTables: # regular tables
dbc.query("select speed from db.{}".format(rTbName[0]))
numbers = dbc.getQueryResult()
for row in numbers :
for row in numbers:
# print("<{}>".format(n), end="", flush=True)
bList.add(row[0])
print("Top numbers in DB right now: {}".format(bList))
print("TDengine client execution is about to start in 2 seconds...")
time.sleep(2.0)
dbManager = None # release?
dbManager = None # release?
def prepare(self):
self._printLastNumbers()
def run(self):
if gConfig.auto_start_service :
if gConfig.auto_start_service:
svcMgr = SvcManager()
svcMgr.startTaosService()
self._printLastNumbers()
dbManager = DbManager() # Regular function
dbManager = DbManager() # Regular function
thPool = ThreadPool(gConfig.num_threads, gConfig.max_steps)
self.tc = ThreadCoordinator(thPool, dbManager)
self.tc.run()
# print("exec stats: {}".format(self.tc.getExecStats()))
# print("TC failed = {}".format(self.tc.isFailed()))
if gConfig.auto_start_service :
# print("TC failed = {}".format(self.tc.isFailed()))
if gConfig.auto_start_service:
svcMgr.stopTaosService()
# Print exec status, etc., AFTER showing messages from the server
self.conclude()
# print("TC failed (2) = {}".format(self.tc.isFailed()))
return 1 if self.tc.isFailed() else 0 # Linux return code: ref https://shapeshed.com/unix-exit-codes/
# Linux return code: ref https://shapeshed.com/unix-exit-codes/
return 1 if self.tc.isFailed() else 0
def conclude(self):
self.tc.printStats()
self.tc.getDbManager().cleanUp()
self.tc.getDbManager().cleanUp()
class MainExec:
STATUS_STARTING = 1
STATUS_RUNNING = 2
STATUS_RUNNING = 2
STATUS_STOPPING = 3
STATUS_STOPPED = 4
STATUS_STOPPED = 4
@classmethod
def runClient(cls):
......@@ -2004,13 +2337,13 @@ class MainExec:
svcManager.run()
@classmethod
def runTemp(cls): # for debugging purposes
def runTemp(cls): # for debugging purposes
# # Hack to exercise reading from disk, imcreasing coverage. TODO: fix
# dbc = dbState.getDbConn()
# sTbName = dbState.getFixedSuperTableName()
# sTbName = dbState.getFixedSuperTableName()
# dbc.execute("create database if not exists db")
# if not dbState.getState().equals(StateEmpty()):
# dbc.execute("use db")
# dbc.execute("use db")
# rTables = None
# try: # the super table may not exist
......@@ -2022,7 +2355,7 @@ class MainExec:
# logger.info("Result: {}".format(rTables))
# except taos.error.ProgrammingError as err:
# logger.info("Initial Super table OPS error: {}".format(err))
# # sys.exit()
# if ( not rTables == None):
# # print("rTables[0] = {}, type = {}".format(rTables[0], type(rTables[0])))
......@@ -2031,24 +2364,26 @@ class MainExec:
# ds = dbState
# logger.info("Inserting into table: {}".format(rTbName[0]))
# sql = "insert into db.{} values ('{}', {});".format(
# rTbName[0],
# rTbName[0],
# ds.getNextTick(), ds.getNextInt())
# dbc.execute(sql)
# for rTbName in rTables : # regular tables
# for rTbName in rTables : # regular tables
# dbc.query("select * from db.{}".format(rTbName[0])) # TODO: check success failure
# logger.info("Initial READING operation is successful")
# logger.info("Initial READING operation is successful")
# except taos.error.ProgrammingError as err:
# logger.info("Initial WRITE/READ error: {}".format(err))
# logger.info("Initial WRITE/READ error: {}".format(err))
# Sandbox testing code
# dbc = dbState.getDbConn()
# while True:
# rows = dbc.query("show databases")
# rows = dbc.query("show databases")
# print("Rows: {}, time={}".format(rows, time.time()))
return
return
def main():
# Super cool Python argument library: https://docs.python.org/3/library/argparse.html
# Super cool Python argument library:
# https://docs.python.org/3/library/argparse.html
parser = argparse.ArgumentParser(
formatter_class=argparse.RawDescriptionHelpFormatter,
description=textwrap.dedent('''\
......@@ -2059,50 +2394,90 @@ def main():
'''))
parser.add_argument('-a', '--auto-start-service', action='store_true',
help='Automatically start/stop the TDengine service (default: false)')
parser.add_argument('-c', '--connector-type', action='store', default='native', type=str,
help='Connector type to use: native, rest, or mixed (default: 10)')
parser.add_argument('-d', '--debug', action='store_true',
help='Turn on DEBUG mode for more logging (default: false)')
parser.add_argument('-e', '--run-tdengine', action='store_true',
help='Run TDengine service in foreground (default: false)')
parser.add_argument('-l', '--larger-data', action='store_true',
help='Write larger amount of data during write operations (default: false)')
parser.add_argument('-p', '--per-thread-db-connection', action='store_true',
help='Use a single shared db connection (default: false)')
parser.add_argument('-r', '--record-ops', action='store_true',
help='Use a pair of always-fsynced fils to record operations performing + performed, for power-off tests (default: false)')
parser.add_argument('-s', '--max-steps', action='store', default=1000, type=int,
help='Maximum number of steps to run (default: 100)')
parser.add_argument('-t', '--num-threads', action='store', default=5, type=int,
help='Number of threads to run (default: 10)')
parser.add_argument(
'-a',
'--auto-start-service',
action='store_true',
help='Automatically start/stop the TDengine service (default: false)')
parser.add_argument(
'-c',
'--connector-type',
action='store',
default='native',
type=str,
help='Connector type to use: native, rest, or mixed (default: 10)')
parser.add_argument(
'-d',
'--debug',
action='store_true',
help='Turn on DEBUG mode for more logging (default: false)')
parser.add_argument(
'-e',
'--run-tdengine',
action='store_true',
help='Run TDengine service in foreground (default: false)')
parser.add_argument(
'-l',
'--larger-data',
action='store_true',
help='Write larger amount of data during write operations (default: false)')
parser.add_argument(
'-p',
'--per-thread-db-connection',
action='store_true',
help='Use a single shared db connection (default: false)')
parser.add_argument(
'-r',
'--record-ops',
action='store_true',
help='Use a pair of always-fsynced fils to record operations performing + performed, for power-off tests (default: false)')
parser.add_argument(
'-s',
'--max-steps',
action='store',
default=1000,
type=int,
help='Maximum number of steps to run (default: 100)')
parser.add_argument(
'-t',
'--num-threads',
action='store',
default=5,
type=int,
help='Number of threads to run (default: 10)')
parser.add_argument(
'-x',
'--continue-on-exception',
action='store_true',
help='Continue execution after encountering unexpected/disallowed errors/exceptions (default: false)')
global gConfig
gConfig = parser.parse_args()
# Logging Stuff
global logger
_logger = logging.getLogger('CrashGen') # real logger
_logger.addFilter(LoggingFilter())
_logger = logging.getLogger('CrashGen') # real logger
_logger.addFilter(LoggingFilter())
ch = logging.StreamHandler()
_logger.addHandler(ch)
logger = MyLoggingAdapter(_logger, []) # Logging adapter, to be used as a logger
# Logging adapter, to be used as a logger
logger = MyLoggingAdapter(_logger, [])
if ( gConfig.debug ):
logger.setLevel(logging.DEBUG) # default seems to be INFO
if (gConfig.debug):
logger.setLevel(logging.DEBUG) # default seems to be INFO
else:
logger.setLevel(logging.INFO)
Dice.seed(0) # initial seeding of dice
Dice.seed(0) # initial seeding of dice
# Run server or client
if gConfig.run_tdengine : # run server
if gConfig.run_tdengine: # run server
MainExec.runService()
else :
else:
return MainExec.runClient()
if __name__ == "__main__":
exitCode = main()
# print("Exiting with code: {}".format(exitCode))
......
tests/pytest/crash_gen.sh
浏览文件 @ 70625808
......@@ -31,11 +31,22 @@ then
exit -1
fi
CURR_DIR=`pwd`
IN_TDINTERNAL="community"
if [[ "$CURR_DIR" == *"$IN_TDINTERNAL"* ]]; then
TAOS_DIR=$CURR_DIR/../../..
else
TAOS_DIR=$CURR_DIR/../..
fi
TAOSD_DIR=`find $TAOS_DIR -name "taosd"|grep bin|head -n1`
LIB_DIR=`echo $TAOSD_DIR|rev|cut -d '/' -f 3,4,5,6|rev`/lib
# First we need to set up a path for Python to find our own TAOS modules, so that "import" can work.
export PYTHONPATH=$(pwd)/../../src/connector/python/linux/python3
# Then let us set up the library path so that our compiled SO file can be loaded by Python
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$(pwd)/../../build/build/lib
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$LIB_DIR
# Now we are all let, and let's see if we can find a crash. Note we pass all params
python3 ./crash_gen.py $@
tests/pytest/fulltest.sh
浏览文件 @ 70625808
......@@ -121,7 +121,7 @@ python3 ./test.py -f import_merge/importTORestart.py
python3 ./test.py -f import_merge/importTPORestart.py
python3 ./test.py -f import_merge/importTRestart.py
python3 ./test.py -f import_merge/importInsertThenImport.py
python3 ./test.py -f import_merge/importCSV.py
# user
python3 ./test.py -f user/user_create.py
python3 ./test.py -f user/pass_len.py
......@@ -156,3 +156,7 @@ python3 ./test.py -f alter/alter_table_crash.py
# client
python3 ./test.py -f client/client.py
# Misc
python3 testCompress.py
python3 testNoCompress.py
tests/pytest/regressiontest.sh
浏览文件 @ 70625808
......@@ -121,7 +121,7 @@ python3 ./test.py -f import_merge/importTORestart.py
python3 ./test.py -f import_merge/importTPORestart.py
python3 ./test.py -f import_merge/importTRestart.py
python3 ./test.py -f import_merge/importInsertThenImport.py
python3 ./test.py -f import_merge/importCSV.py
# user
python3 ./test.py -f user/user_create.py
python3 ./test.py -f user/pass_len.py
......@@ -150,3 +150,7 @@ python3 ./test.py -f alter/alter_table_crash.py
# client
python3 ./test.py -f client/client.py
# Misc
python3 testCompress.py
python3 testNoCompress.py
tests/pytest/util/dnodes-random-fail.py 0 → 100644
浏览文件 @ 70625808
###################################################################
# Copyright (c) 2016 by TAOS Technologies, Inc.
# All rights reserved.
#
# This file is proprietary and confidential to TAOS Technologies.
# No part of this file may be reproduced, stored, transmitted,
# disclosed or used in any form or by any means other than as
# expressly provided by the written permission from Jianhui Tao
#
###################################################################
# -*- coding: utf-8 -*-
import sys
import os
import os.path
import subprocess
from util.log import *
class TDSimClient:
def __init__(self):
self.testCluster = False
self.cfgDict = {
"numOfLogLines": "100000000",
"numOfThreadsPerCore": "2.0",
"locale": "en_US.UTF-8",
"charset": "UTF-8",
"asyncLog": "0",
"anyIp": "0",
"sdbDebugFlag": "135",
"rpcDebugFlag": "135",
"tmrDebugFlag": "131",
"cDebugFlag": "135",
"udebugFlag": "135",
"jnidebugFlag": "135",
"qdebugFlag": "135",
}
def init(self, path):
self.__init__()
self.path = path
def getLogDir(self):
self.logDir = "%s/sim/psim/log" % (self.path)
return self.logDir
def getCfgDir(self):
self.cfgDir = "%s/sim/psim/cfg" % (self.path)
return self.cfgDir
def setTestCluster(self, value):
self.testCluster = value
def addExtraCfg(self, option, value):
self.cfgDict.update({option: value})
def cfg(self, option, value):
cmd = "echo '%s %s' >> %s" % (option, value, self.cfgPath)
if os.system(cmd) != 0:
tdLog.exit(cmd)
def deploy(self):
self.logDir = "%s/sim/psim/log" % (self.path)
self.cfgDir = "%s/sim/psim/cfg" % (self.path)
self.cfgPath = "%s/sim/psim/cfg/taos.cfg" % (self.path)
cmd = "rm -rf " + self.logDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "mkdir -p " + self.logDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "rm -rf " + self.cfgDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "mkdir -p " + self.cfgDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "touch " + self.cfgPath
if os.system(cmd) != 0:
tdLog.exit(cmd)
if self.testCluster:
self.cfg("masterIp", "192.168.0.1")
self.cfg("secondIp", "192.168.0.2")
self.cfg("logDir", self.logDir)
for key, value in self.cfgDict.items():
self.cfg(key, value)
tdLog.debug("psim is deployed and configured by %s" % (self.cfgPath))
class TDDnode:
def __init__(self, index):
self.index = index
self.running = 0
self.deployed = 0
self.testCluster = False
self.valgrind = 0
def init(self, path):
self.path = path
def setTestCluster(self, value):
self.testCluster = value
def setValgrind(self, value):
self.valgrind = value
def getDataSize(self):
totalSize = 0
if (self.deployed == 1):
for dirpath, dirnames, filenames in os.walk(self.dataDir):
for f in filenames:
fp = os.path.join(dirpath, f)
if not os.path.islink(fp):
totalSize = totalSize + os.path.getsize(fp)
return totalSize
def deploy(self):
self.logDir = "%s/sim/dnode%d/log" % (self.path, self.index)
self.dataDir = "%s/sim/dnode%d/data" % (self.path, self.index)
self.cfgDir = "%s/sim/dnode%d/cfg" % (self.path, self.index)
self.cfgPath = "%s/sim/dnode%d/cfg/taos.cfg" % (
self.path, self.index)
cmd = "rm -rf " + self.dataDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "rm -rf " + self.logDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "rm -rf " + self.cfgDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "mkdir -p " + self.dataDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "mkdir -p " + self.logDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "mkdir -p " + self.cfgDir
if os.system(cmd) != 0:
tdLog.exit(cmd)
cmd = "touch " + self.cfgPath
if os.system(cmd) != 0:
tdLog.exit(cmd)
if self.testCluster:
self.startIP()
if self.testCluster:
self.cfg("masterIp", "192.168.0.1")
self.cfg("secondIp", "192.168.0.2")
self.cfg("publicIp", "192.168.0.%d" % (self.index))
self.cfg("internalIp", "192.168.0.%d" % (self.index))
self.cfg("privateIp", "192.168.0.%d" % (self.index))
self.cfg("dataDir", self.dataDir)
self.cfg("logDir", self.logDir)
self.cfg("numOfLogLines", "100000000")
self.cfg("mnodeEqualVnodeNum", "0")
self.cfg("walLevel", "1")
self.cfg("statusInterval", "1")
self.cfg("numOfTotalVnodes", "64")
self.cfg("numOfMnodes", "3")
self.cfg("numOfThreadsPerCore", "2.0")
self.cfg("monitor", "0")
self.cfg("maxVnodeConnections", "30000")
self.cfg("maxMgmtConnections", "30000")
self.cfg("maxMeterConnections", "30000")
self.cfg("maxShellConns", "30000")
self.cfg("locale", "en_US.UTF-8")
self.cfg("charset", "UTF-8")
self.cfg("asyncLog", "0")
self.cfg("anyIp", "0")
self.cfg("dDebugFlag", "135")
self.cfg("mDebugFlag", "135")
self.cfg("sdbDebugFlag", "135")
self.cfg("rpcDebugFlag", "135")
self.cfg("tmrDebugFlag", "131")
self.cfg("cDebugFlag", "135")
self.cfg("httpDebugFlag", "135")
self.cfg("monitorDebugFlag", "135")
self.cfg("udebugFlag", "135")
self.cfg("jnidebugFlag", "135")
self.cfg("qdebugFlag", "135")
self.deployed = 1
tdLog.debug(
"dnode:%d is deployed and configured by %s" %
(self.index, self.cfgPath))
def getBuildPath(self):
selfPath = os.path.dirname(os.path.realpath(__file__))
if ("community" in selfPath):
projPath = selfPath[:selfPath.find("community")]
else:
projPath = selfPath[:selfPath.find("tests")]
for root, dirs, files in os.walk(projPath):
if ("taosd" in files):
rootRealPath = os.path.dirname(os.path.realpath(root))
if ("packaging" not in rootRealPath):
buildPath = root[:len(root)-len("/build/bin")]
break
return buildPath
def start(self):
buildPath = self.getBuildPath()
if (buildPath == ""):
tdLog.exit("taosd not found!")
else:
tdLog.info("taosd found in %s" % buildPath)
binPath = buildPath + "/build/bin/taosd"
if self.deployed == 0:
tdLog.exit("dnode:%d is not deployed" % (self.index))
if self.valgrind == 0:
cmd = "nohup %s -c %s > /dev/null 2>&1 & " % (
binPath, self.cfgDir)
else:
valgrindCmdline = "valgrind --tool=memcheck --leak-check=full --show-reachable=no --track-origins=yes --show-leak-kinds=all -v --workaround-gcc296-bugs=yes"
cmd = "nohup %s %s -c %s --random-file-fail-factor 5 2>&1 & " % (
valgrindCmdline, binPath, self.cfgDir)
print(cmd)
if os.system(cmd) != 0:
tdLog.exit(cmd)
self.running = 1
tdLog.debug("dnode:%d is running with %s " % (self.index, cmd))
tdLog.debug("wait 5 seconds for the dnode:%d to start." % (self.index))
time.sleep(5)
def stop(self):
if self.valgrind == 0:
toBeKilled = "taosd"
else:
toBeKilled = "valgrind.bin"
if self.running != 0:
psCmd = "ps -ef|grep -w %s| grep -v grep | awk '{print $2}'" % toBeKilled
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
while(processID):
killCmd = "kill -INT %s > /dev/null 2>&1" % processID
os.system(killCmd)
time.sleep(1)
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
for port in range(6030, 6041):
fuserCmd = "fuser -k -n tcp %d" % port
os.system(fuserCmd)
if self.valgrind:
time.sleep(2)
self.running = 0
tdLog.debug("dnode:%d is stopped by kill -INT" % (self.index))
def forcestop(self):
if self.valgrind == 0:
toBeKilled = "taosd"
else:
toBeKilled = "valgrind.bin"
if self.running != 0:
psCmd = "ps -ef|grep -w %s| grep -v grep | awk '{print $2}'" % toBeKilled
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
while(processID):
killCmd = "kill -KILL %s > /dev/null 2>&1" % processID
os.system(killCmd)
time.sleep(1)
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
for port in range(6030, 6041):
fuserCmd = "fuser -k -n tcp %d" % port
os.system(fuserCmd)
if self.valgrind:
time.sleep(2)
self.running = 0
tdLog.debug("dnode:%d is stopped by kill -KILL" % (self.index))
def startIP(self):
cmd = "sudo ifconfig lo:%d 192.168.0.%d up" % (self.index, self.index)
if os.system(cmd) != 0:
tdLog.exit(cmd)
def stopIP(self):
cmd = "sudo ifconfig lo:%d 192.168.0.%d down" % (
self.index, self.index)
if os.system(cmd) != 0:
tdLog.exit(cmd)
def cfg(self, option, value):
cmd = "echo '%s %s' >> %s" % (option, value, self.cfgPath)
if os.system(cmd) != 0:
tdLog.exit(cmd)
def getDnodeRootDir(self, index):
dnodeRootDir = "%s/sim/psim/dnode%d" % (self.path, index)
return dnodeRootDir
def getDnodesRootDir(self):
dnodesRootDir = "%s/sim/psim" % (self.path)
return dnodesRootDir
class TDDnodes:
def __init__(self):
self.dnodes = []
self.dnodes.append(TDDnode(1))
self.dnodes.append(TDDnode(2))
self.dnodes.append(TDDnode(3))
self.dnodes.append(TDDnode(4))
self.dnodes.append(TDDnode(5))
self.dnodes.append(TDDnode(6))
self.dnodes.append(TDDnode(7))
self.dnodes.append(TDDnode(8))
self.dnodes.append(TDDnode(9))
self.dnodes.append(TDDnode(10))
self.simDeployed = False
def init(self, path):
psCmd = "ps -ef|grep -w taosd| grep -v grep | awk '{print $2}'"
processID = subprocess.check_output(psCmd, shell=True).decode("utf-8")
while(processID):
killCmd = "kill -KILL %s > /dev/null 2>&1" % processID
os.system(killCmd)
time.sleep(1)
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
psCmd = "ps -ef|grep -w valgrind.bin| grep -v grep | awk '{print $2}'"
processID = subprocess.check_output(psCmd, shell=True).decode("utf-8")
while(processID):
killCmd = "kill -KILL %s > /dev/null 2>&1" % processID
os.system(killCmd)
time.sleep(1)
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
binPath = os.path.dirname(os.path.realpath(__file__))
binPath = binPath + "/../../../debug/"
tdLog.debug("binPath %s" % (binPath))
binPath = os.path.realpath(binPath)
tdLog.debug("binPath real path %s" % (binPath))
# cmd = "sudo cp %s/build/lib/libtaos.so /usr/local/lib/taos/" % (binPath)
# tdLog.debug(cmd)
# os.system(cmd)
# cmd = "sudo cp %s/build/bin/taos /usr/local/bin/taos/" % (binPath)
# if os.system(cmd) != 0 :
# tdLog.exit(cmd)
# tdLog.debug("execute %s" % (cmd))
# cmd = "sudo cp %s/build/bin/taosd /usr/local/bin/taos/" % (binPath)
# if os.system(cmd) != 0 :
# tdLog.exit(cmd)
# tdLog.debug("execute %s" % (cmd))
if path == "":
# self.path = os.path.expanduser('~')
self.path = os.path.abspath(binPath + "../../")
else:
self.path = os.path.realpath(path)
for i in range(len(self.dnodes)):
self.dnodes[i].init(self.path)
self.sim = TDSimClient()
self.sim.init(self.path)
def setTestCluster(self, value):
self.testCluster = value
def setValgrind(self, value):
self.valgrind = value
def deploy(self, index):
self.sim.setTestCluster(self.testCluster)
if (self.simDeployed == False):
self.sim.deploy()
self.simDeployed = True
self.check(index)
self.dnodes[index - 1].setTestCluster(self.testCluster)
self.dnodes[index - 1].setValgrind(self.valgrind)
self.dnodes[index - 1].deploy()
def cfg(self, index, option, value):
self.check(index)
self.dnodes[index - 1].cfg(option, value)
def start(self, index):
self.check(index)
self.dnodes[index - 1].start()
def stop(self, index):
self.check(index)
self.dnodes[index - 1].stop()
def getDataSize(self, index):
self.check(index)
return self.dnodes[index - 1].getDataSize()
def forcestop(self, index):
self.check(index)
self.dnodes[index - 1].forcestop()
def startIP(self, index):
self.check(index)
if self.testCluster:
self.dnodes[index - 1].startIP()
def stopIP(self, index):
self.check(index)
if self.dnodes[index - 1].testCluster:
self.dnodes[index - 1].stopIP()
def check(self, index):
if index < 1 or index > 10:
tdLog.exit("index:%d should on a scale of [1, 10]" % (index))
def stopAll(self):
tdLog.info("stop all dnodes")
for i in range(len(self.dnodes)):
self.dnodes[i].stop()
psCmd = "ps -ef | grep -w taosd | grep 'root' | grep -v grep | awk '{print $2}'"
processID = subprocess.check_output(psCmd, shell=True).decode("utf-8")
if processID:
cmd = "sudo systemctl stop taosd"
os.system(cmd)
# if os.system(cmd) != 0 :
# tdLog.exit(cmd)
psCmd = "ps -ef|grep -w taosd| grep -v grep | awk '{print $2}'"
processID = subprocess.check_output(psCmd, shell=True).decode("utf-8")
while(processID):
killCmd = "kill -KILL %s > /dev/null 2>&1" % processID
os.system(killCmd)
time.sleep(1)
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
psCmd = "ps -ef|grep -w valgrind.bin| grep -v grep | awk '{print $2}'"
processID = subprocess.check_output(psCmd, shell=True).decode("utf-8")
while(processID):
killCmd = "kill -KILL %s > /dev/null 2>&1" % processID
os.system(killCmd)
time.sleep(1)
processID = subprocess.check_output(
psCmd, shell=True).decode("utf-8")
# if os.system(cmd) != 0 :
# tdLog.exit(cmd)
def getDnodesRootDir(self):
dnodesRootDir = "%s/sim" % (self.path)
return dnodesRootDir
def getSimCfgPath(self):
return self.sim.getCfgDir()
def getSimLogPath(self):
return self.sim.getLogDir()
def addSimExtraCfg(self, option, value):
self.sim.addExtraCfg(option, value)
tdDnodes = TDDnodes()
tests/script/jenkins/basic.txt
浏览文件 @ 70625808
......@@ -117,8 +117,6 @@ cd ../../../debug; make
./test.sh -f general/parser/import_commit3.sim
./test.sh -f general/parser/insert_tb.sim
./test.sh -f general/parser/first_last.sim
# dyh is processing this script
#./test.sh -f general/parser/import_file.sim
./test.sh -f general/parser/lastrow.sim
./test.sh -f general/parser/nchar.sim
./test.sh -f general/parser/null_char.sim
......@@ -145,7 +143,6 @@ cd ../../../debug; make
./test.sh -f general/parser/groupby.sim
./test.sh -f general/parser/set_tag_vals.sim
#./test.sh -f general/parser/sliding.sim
./test.sh -f general/parser/tags_dynamically_specifiy.sim
./test.sh -f general/parser/tags_filter.sim
./test.sh -f general/parser/slimit_alter_tags.sim
./test.sh -f general/parser/join.sim
......
tests/script/sh/deploy.sh
浏览文件 @ 70625808
......@@ -125,7 +125,6 @@ echo "mqttDebugFlag 131" >> $TAOS_CFG
echo "qdebugFlag 135" >> $TAOS_CFG
echo "rpcDebugFlag 135" >> $TAOS_CFG
echo "tmrDebugFlag 131" >> $TAOS_CFG
echo "cDebugFlag 135" >> $TAOS_CFG
echo "udebugFlag 135" >> $TAOS_CFG
echo "sdebugFlag 135" >> $TAOS_CFG
echo "wdebugFlag 135" >> $TAOS_CFG
......
tests/script/sh/exec-random-fail.sh 0 → 100755
浏览文件 @ 70625808
#!/bin/bash
# if [ $# != 4 || $# != 5 ]; then
# echo "argument list need input : "
# echo " -n nodeName"
# echo " -s start/stop"
# echo " -c clear"
# exit 1
# fi
NODE_NAME=
EXEC_OPTON=
CLEAR_OPTION="false"
while getopts "n:s:u:x:ct" arg
do
case $arg in
n)
NODE_NAME=$OPTARG
;;
s)
EXEC_OPTON=$OPTARG
;;
c)
CLEAR_OPTION="clear"
;;
t)
SHELL_OPTION="true"
;;
u)
USERS=$OPTARG
;;
x)
SIGNAL=$OPTARG
;;
?)
echo "unkown argument"
;;
esac
done
SCRIPT_DIR=`dirname $0`
cd $SCRIPT_DIR/../
SCRIPT_DIR=`pwd`
IN_TDINTERNAL="community"
if [[ "$SCRIPT_DIR" == *"$IN_TDINTERNAL"* ]]; then
cd ../../..
else
cd ../../
fi
TAOS_DIR=`pwd`
TAOSD_DIR=`find . -name "taosd"|grep bin|head -n1`
if [[ "$TAOSD_DIR" == *"$IN_TDINTERNAL"* ]]; then
BIN_DIR=`find . -name "taosd"|grep bin|head -n1|cut -d '/' --fields=2,3`
else
BIN_DIR=`find . -name "taosd"|grep bin|head -n1|cut -d '/' --fields=2`
fi
BUILD_DIR=$TAOS_DIR/$BIN_DIR/build
SIM_DIR=$TAOS_DIR/sim
NODE_DIR=$SIM_DIR/$NODE_NAME
EXE_DIR=$BUILD_DIR/bin
CFG_DIR=$NODE_DIR/cfg
LOG_DIR=$NODE_DIR/log
DATA_DIR=$NODE_DIR/data
MGMT_DIR=$NODE_DIR/data/mgmt
TSDB_DIR=$NODE_DIR/data/tsdb
TAOS_CFG=$NODE_DIR/cfg/taos.cfg
echo ------------ $EXEC_OPTON $NODE_NAME
TAOS_FLAG=$SIM_DIR/tsim/flag
if [ -f "$TAOS_FLAG" ]; then
EXE_DIR=/usr/local/bin/taos
fi
if [ "$CLEAR_OPTION" = "clear" ]; then
echo rm -rf $MGMT_DIR $TSDB_DIR
rm -rf $TSDB_DIR
rm -rf $MGMT_DIR
fi
if [ "$EXEC_OPTON" = "start" ]; then
echo "ExcuteCmd:" $EXE_DIR/taosd -c $CFG_DIR
if [ "$SHELL_OPTION" = "true" ]; then
nohup valgrind --log-file=${LOG_DIR}/valgrind.log --tool=memcheck --leak-check=full --show-reachable=no --track-origins=yes --show-leak-kinds=all -v --workaround-gcc296-bugs=yes $EXE_DIR/taosd -c $CFG_DIR > /dev/null 2>&1 &
else
nohup $EXE_DIR/taosd -c $CFG_DIR --random-file-fail-factor 5 > /dev/null 2>&1 &
fi
else
#relative path
RCFG_DIR=sim/$NODE_NAME/cfg
PID=`ps -ef|grep taosd | grep $RCFG_DIR | grep -v grep | awk '{print $2}'`
while [ -n "$PID" ]
do
if [ "$SIGNAL" = "SIGINT" ]; then
echo try to kill by signal SIGINT
kill -SIGINT $PID
else
echo try to kill by signal SIGKILL
kill -9 $PID
fi
sleep 1
PID=`ps -ef|grep taosd | grep $RCFG_DIR | grep -v grep | awk '{print $2}'`
done
fi
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或