- 19 5月, 2017 4 次提交
-
-
由 Pengzhou Tang 提交于
Resource group cpu rate limitation is implemented with cgroup on linux system. When resource group is enabled via GUC we check whether cgroup is available and properly configured on the system. A sub cgroup is created for each resource group, cpu quota and share weight will be set depends on the resource group configuration. The queries will run under these cgroups, and the cpu usage will be restricted by cgroup. The cgroups directory structures: * /sys/fs/cgroup/{cpu,cpuacct}/gpdb: the toplevel gpdb cgroup * /sys/fs/cgroup/{cpu,cpuacct}/gpdb/*/: cgroup for each resource group The logic for cpu rate limitation: * in toplevel gpdb cgroup we set the cpu quota and share weight as: cpu.cfs_quota_us := cpu.cfs_period_us * 256 * gp_resource_group_cpu_limit cpu.shares := 1024 * ncores * for each sub group we set the cpu quota and share weight as: sub.cpu.cfs_quota_us := -1 sub.cpu.shares := top.cpu.shares * sub.cpu_rate_limit The minimum and maximum cpu percentage for a sub cgroup: sub.cpu.min_percentage := gp_resource_group_cpu_limit * sub.cpu_rate_limit sub.cpu.max_percentage := gp_resource_group_cpu_limit The acutal percentage depends on how busy the system is. gp_resource_group_cpu_limit is a GUC introduced to control the cpu resgroups assigned on each host. gpconfig -c gp_resource_group_cpu_limit -v '0.9' A new pipeline is created to perform the tests as we need privileged permission to enable and setup cgroups on the system. Signed-off-by: NNing Yu <nyu@pivotal.io>
-
由 Venkatesh Raghavan 提交于
In the updated tests, we used functions like disable_xform and enable_xform to hint the optimizer to disallow/allow a particular physical node. However, these functions are only available when GPDB is built with GPORCA. Planner on the other hand accomplished this via a GUC. To avoid usage of these functions in tests, I have introduced couple of GUCS that mimic the same planner behavior but now for GPORCA. In this effort I needed to add an API inside GPORCA.
-
由 Ashwin Agrawal 提交于
-
由 Larry Hamel 提交于
Please note: this test does not pass on macOS because of an apparent failure of libsigar. See comment in the test on how to omit a test on MacOS. Signed-off-by: NMelanie Plageman <mplageman@pivotal.io> Signed-off-by: NLarry Hamel <lhamel@pivotal.io>
-
- 18 5月, 2017 6 次提交
-
-
由 Marbin Tan 提交于
* gpperfmon: update gpperfmon install user docs IPv6 Follow up for commit c52c768e [ci-skip]
-
由 Larry Hamel 提交于
Signed-off-by: NMarbin Tan <mtan@pivotal.io>
-
由 Daniel Gustafsson 提交于
When dumping a table with dropped columns, the columns are created and dropped again during the restore to create binary-compatible heap files. Using TEXT as the dummy attribute however cause toast tables to be created, even for tables which have no other toastable attributes. This cause catalog bloat and cause problems for upgrade since the newly created toast table doesn't have a synchronized Oid in the dumpfile. Fix by using INTEGER instead. This is a partial backport of upstream commit 53fa850c which fixes this case and more (where the more part relies on code Greenplum has yet to merge, thus partial backport). commit 53fa850c Author: Tom Lane <tgl@sss.pgh.pa.us> Date: Thu Jul 2 21:34:32 2009 +0000 Fix up pg_dump's --binary-upgrade option so that it behaves properly with inherited columns and check constraints. Per my recent trouble report.
-
由 Daniel Gustafsson 提交于
Defer creation of the PQExpBuffer until we know that the object is in fact dumpable to avoid leaking a few buffers.
-
由 Daniel Gustafsson 提交于
The opfamily dump in pg_dump had an unfortunate typo that snuck in with the initial pg_upgrade commit, and pg_dump had a broken check query. Fix syntax.
-
由 Daniel Gustafsson 提交于
Previously pg_upgrade was unable to upgrade from a 5.0 cluster due to Text Search objects, enums, extensions and VIEW rules were left out from the original import of pg_upgrade. This adds support for these objects. Further, the binary-upgrade mode in pg_dump didn't synchronize the Oids for the functions associated with the language. Extract and add the Oids into the dump to fix. Also expand comments and fix a missing static declaration function prototype.
-
- 17 5月, 2017 10 次提交
-
-
由 Lisa Owen 提交于
-
由 Chuck Litzell 提交于
-
由 Heikki Linnakangas 提交于
Fixes github issue #1774.
-
由 mkiyama 提交于
-
由 Chris Hajas 提交于
Since python 2.7 has the unittest module included, we do not need to use unittest2 anymore. This removes this dependency requirement. We previously used the unittest2 module as it included features that were not present in python 2.6. There are still unittest2 dependencies in the TINC tests, but that may be a more intensive effort.
-
由 C.J. Jameson 提交于
* Fortify SED_PG_CONF() shell function, as called by gpinitsystem It's also called via gpcreateseg.sh There is a chance where gpinitsystem may miss a GUC change in one of the segments and left undetected. This would cause failures down the line. More robust in three ways: - Retry ssh calls in case they fail - Actually error exit if it does fail, for quicker diagnosis if the value in postgres.conf did not get updated - When keeping a version of the key-value pair in a comment, only archive the first-matched key-value (otherwise you get comments on your comments) Use trap for retrying failed ssh Use trap in a limited scope for now
-
由 Marbin Tan 提交于
If IPv6 is enabled, gpperfmon will complain that it can't connect to the database because, gpmon does not have permission to access through ::1. - Add behave test
-
由 Marbin Tan 提交于
-
由 Todd Sedano 提交于
- We only want to copy the file if it exists
-
由 Michael Roth 提交于
Updates to configure to use `getversion --short` to generate a version tag that can be used in `Makefile.global` for packaging etc...
-
- 16 5月, 2017 15 次提交
-
-
由 Lisa Owen 提交于
* add info about pl/python multidim array support * restructure array info so better matches upstream docs * reorganize a bit, davids other requested edits
-
由 Daniel Gustafsson 提交于
In the past there was functionality to decorate a stacktrace with the developer in charge of the separate functions. This is not in use anymore and it's not something we want either.
-
由 Haisheng Yuan 提交于
The following 2 functions are removed: * ExecShareInputScanMarkPos * ExecShareInputScanRestrPos
-
由 Todd Sedano 提交于
Signed-off-by: NChris Hajas <chajas@pivotal.io>
-
由 Jesse Zhang 提交于
-
由 Michael Roth 提交于
GP_VERSION is used by gppkg and other bits to show what version of GPDB it is build from. This is also propogated via gp_config for all packages and should have a valid value for GP_VERSION.
-
由 Karen Huddleston 提交于
- We haven't changed cdb_bsa_util for quite some time and these tests are not useful to us. Additionally, changing the Netbackup version we compiled them with caused compilation to fail.
-
由 Jamie McAtamney 提交于
Signed-off-by: NKaren Huddleston <khuddleston@pivotal.io>
-
由 C.J. Jameson 提交于
- Make sure to create the database first and then run the sql Signed-off-by: NMarbin Tan <mtan@pivotal.io>
-
由 Marbin Tan 提交于
Across versions, separate files had been created. Going forward, there will be just one gpperfmon.sql for setting up the gpperfmon database. - Remove other iterations of gpperfmon sql in makefile Signed-off-by: NC.J. Jameson <cjameson@pivotal.io>
-
由 Jesse Zhang 提交于
-
由 Lisa Owen 提交于
* update docs for functions supporting COST and DEFAULT arg clauses * updates per davids edits
-
由 Jane Beckman 提交于
-
由 Chuck Litzell 提交于
-
- 15 5月, 2017 3 次提交
-
-
由 Venkatesh Raghavan 提交于
* Enable analyzing root partitions * Ensure that the name of the guc is clear * Remove double negation (where possible) * Update comments * Co-locate gucs that have similar purpose * Remove dead gucs * Classify them correctly so that they are no longer hidden
-
由 Heikki Linnakangas 提交于
We had cherry-picked some of the PostgreSQL 9.3 functions, but not all the subsequent fixes to them. I believe we need at least upstream commit 66008564 to fix github issue #2430, and surely the other bug fixes are needed too. Rather than cherry-pick individual commits, this replaces the JSON related source files with the upstream files in toto. This will make diffing and maintenance of these backported JSON function easier in the future.
-
由 Pengzhou Tang 提交于
Formally, GPDB assumed Gp_interconnect_queue_depth was constant during the interconnect life-time which was incorrect for Cursors, if Gp_interconnect_queue_depth was changed after a Cursor was declared, a panic occurred. To avoid this, we make a copy of Gp_interconnect_queue_depth when interconnect is set up. Gp_interconnect_snd_queue_depth has no such problem because it is only used by senders and senders of Cursor will never receive the GUC change command.
-
- 13 5月, 2017 2 次提交
-
-
cache.
-
由 Chuck Litzell 提交于
* Relocates gpperfmon reference from GPCC * Remove emcconnect history and itorators from PDF ditamap * Remove iterator metrics subtopic from PDF ditamap * Additional edits to acknowledge gpperfmon without GPCC * Incorporate gpperfmon.conf reference into gpperfmon_install reference
-