- 29 5月, 2019 19 次提交
-
-
由 xiong-gang 提交于
In commit b43629, we disable two-phase commit if the transaction doesn't write any WAL on the segment. That makes some tests flaky because a read-only transaction may also write WALs on segment and that behavior is indeterminate.
-
由 Shaoqi Bai 提交于
When TARGET_OS is sles, it goes to case *) and return without popd, so leave a dangled pushd, that results later pushd can not work. Co-authored-by: NBradford D. Boyle <bboyle@pivotal.io> Co-authored-by: NShaoqi Bai <sbai@pivotal.io> Reviewed-by: NNing Yu <nyu@pivotal.io>
-
由 Bradford D. Boyle 提交于
Authored-by: NBradford D. Boyle <bboyle@pivotal.io>
-
由 Tingfang Bao 提交于
gp-releng build the libs artifacts for gpdb6, and copy them into gpdb libs when compile the gpdb binary. We found that the libs copied multiple times. eg. ``` [root@a5e93b7cdc29 lib]# ls /usr/local/greenplum-db/lib -l | grep libquicklz -rwxr-xr-x 1 root root 8648 May 8 15:26 libquicklz.so -rwxr-xr-x 1 root root 8648 May 8 15:26 libquicklz.so.1 -rwxr-xr-x 1 root root 8648 May 8 15:26 libquicklz.so.1.5.0 ``` stories: https://www.pivotaltracker.com/story/show/166200423 https://www.pivotaltracker.com/story/show/166200398 https://www.pivotaltracker.com/story/show/166200397Authored-by: NBob Bao <bbao@pivotal.io>
-
由 Bradford D. Boyle 提交于
Co-authored-by: NBradford D. Boyle <bboyle@pivotal.io> Co-authored-by: NJesse Zhang <jzhang@pivotal.io>
-
由 Bradford D. Boyle 提交于
Co-authored-by: NBradford D. Boyle <bboyle@pivotal.io> Co-authored-by: NJesse Zhang <jzhang@pivotal.io>
-
由 Bradford D. Boyle 提交于
Inline tasks are hard to test in isolation because you cannot `fly execute` them. Co-authored-by: NBradford D. Boyle <bboyle@pivotal.io> Co-authored-by: NJesse Zhang <jzhang@pivotal.io>
-
由 Jesse Zhang 提交于
- update release candidate job to produce release candidate artifacts for Ubuntu 18.04 Co-authored-by: NJesse Zhang <jzhang@pivotal.io> Co-authored-by: NBradford D. Boyle <bboyle@pivotal.io>
-
由 Jesse Zhang 提交于
- ubuntu jobs are using the systems compiler and the Docker build image does not contain the GCC 6.4.0 package; CI scripts are modified to conditionally source the gcc environment setup script - when running `make demo-cluster` on ubuntu, it looks like `su` does not preserve the environment set by "greenplum_path.sh" so we re-source "greenplum_path.sh" as gpadmin before starting the demo cluster Co-authored-by: NJesse Zhang <jzhang@pivotal.io> Co-authored-by: NBradford D. Boyle <bboyle@pivotal.io>
-
由 Bradford D. Boyle 提交于
- add ubuntu18.04 as a choice for OS type - include ubuntu18.04 as an OS type for prod pipelines - python, sigar, and zstd build dependencies are included in the gpdb6-ubuntu18.04-build image - the ubuntu docker image has sigar and python installed already and there is no libquicklz-devel package for ubuntu so these inputs have been marked optional in the task - the configure flags were copied from the RHEL7 configure flags - update functions `include_quicklz` and `include_zstd` to handle differences in where the libraries are installed on RHEL7 and Ubuntu platforms - change mv to cp to make local iteration easier; this might also be slightly more performant on Concourse w/ the overlay volumes Co-authored-by: NGoutam Tadi <gtadi@pivotal.io> Co-authored-by: NBradford D. Boyle <bboyle@pivotal.io> Co-authored-by: NJesse Zhang <jzhang@pivotal.io>
-
由 Mark Sliva 提交于
Also add new tests for the integer with units GUC type. Co-authored-by: NKalen Krempely <kkrempely@pivotal.io>
-
由 Mark Sliva 提交于
Escaping double quotes is no longer necessary since we now wrap the entire value in single quotes and correctly escape it before writing it to the postgresql.conf file. Co-authored-by: NJacob Champion <pchampion@pivotal.io>
-
由 Mark Sliva 提交于
Document that when running gpconfig, the value is escaped for users before being written to the postgresql.conf files. Co-authored-by: NJacob Champion <pchampion@pivotal.io>
-
由 Jacob Champion 提交于
Before we were comparing the live value with the quoted value in the file, so it would incorrectly print a GUCS ARE OUT OF SYNC warning. To fix this we had to re-implement the GUC_scanstr function in Python. Co-authored-by: NMark Sliva <msliva@pivotal.io>
-
由 Mark Sliva 提交于
If we don't do this, the postgresql.conf file can easily be corrupted. Co-authored-by: NJacob Champion <pchampion@pivotal.io>
-
由 Jacob Champion 提交于
application_name explicitly filters non-ASCII bytes, so we went with search_path. We had to work around some ASCII/Unicode mixing in the tests, which will probably come back to haunt us during the Python 3 move. Co-authored-by: NMark Sliva <msliva@pivotal.io>
-
由 Jacob Champion 提交于
Currently, gpconfig -c guc -v "''" and gpconfig -c guc -v "''''" result in the same value being written to postgresql.conf: guc='''''' To avoid this confusing behavior, assume that the string passed to gpconfig is the exact value that the user wants to be reflected in the configuration, even if that value already contains single quotes. Co-authored-by: NMark Sliva <msliva@pivotal.io>
-
由 Jacob Champion 提交于
Change to `gpstop -u` for a quick GUC reload as opposed to a full restart during tests. This requires us to choose GUCs that can be reloaded during a SIGHUP (e.g. change gp_resource_group_cpu_limit to checkpoint_completion_target). Co-authored-by: NMark Sliva <msliva@pivotal.io>
-
由 Jacob Champion 提交于
There are no "edge cases". This change requires us to correctly escape the search pattern for all cases, which is desirable anyway. Co-authored-by: NMark Sliva <msliva@pivotal.io>
-
- 28 5月, 2019 6 次提交
-
-
由 Daniel Gustafsson 提交于
The gp_bash_functions Bash library had accumulated some dead code variables, left over from earlier refactorings and removal of utils. Remove and perform a little bit of cleaning. Reviewed-by: Jacob Champion
-
由 xiong-gang 提交于
This bug cause the test 'generate_ao_xlog' hang sometimes. If the current page cannot hold the remaining content (xlp_rem_len) of the record, find the next record in the next page.
-
由 xiong-gang 提交于
'MySessionState' is destroyed before 'groupWaitCancel' is called when shutdown. Access the member of 'MySessionState' in 'groupWaitCancel' may cause segment fault.
-
由 xiong-gang 提交于
Currently, explicit 'BEGIN' creates a full-size writer gang and starts a transaction on it, the following 'END' will commit the transaction in a two-phase way. It can be optimized for some cases: case 1: BEGIN; SELECT * FROM pg_class; END; case 2: BEGIN; SELECT * FROM foo; SELECT * FROM bar; END; case 3: BEGIN; INSERT INTO foo VALUES(1); INSERT INTO bar VALUES(2); END; For case 1, it's unnecessary to create a gang and no need to have two-phase commit. For case 2, it's unnecessary to have two-phase commit as the executors don't write any XLOG. For case 3, don't have to create a full-size writer gang and do two-phase commit on a full-size gang. Co-authored-by: NJialun Du <jdu@pivotal.io>
-
由 Xiaodong Huo 提交于
-
由 Hubert Zhang 提交于
Check hook of string GUC should use char ** instead of const char ** Also remove unused function gpvars_check_memory_spill_ratio() etc.
-
- 27 5月, 2019 6 次提交
-
-
由 Ning Yu 提交于
The backend could set extra data on a vert, such as the raw data collected by the UDF. Also let GDD return each newly created vert so the backend has a chance to set the data.
-
由 Ning Yu 提交于
These dump functions are not part of the GDD algorithm, and they access the backend information, move them from gdddetector.c to gddbackend.c, so gdddetector.c is more focused on the algorithm.
-
由 Ning Yu 提交于
gdddetector.c contains the core GDD algorithm, we used to put the structs in it to hide the implementation details, the purpose was to simplify the detector's usage by providing a small set of carefully designed APIs. However some non-algorithm-related helpers, such as the dump functions, have to be put into this source file to access the struct internals, which increase the effort to understand and maintain the algorithm. So now we move the structs to gdddetectorpriv.h, the name still suggests that it is for private use, however when necessary the callers could access the struct internals by including this header.
-
由 Jimmy Yih 提交于
This commit brings back a scenario from an old, now deleted TINC test, in the form of an isolation2 spec. The basic idea is to switch WAL on master to a new segment while basebackup is in progress. Co-authored-by: NAsim R P <apraveen@pivotal.io>
-
由 Asim R P 提交于
PGAPPNAME is set to "pg_regress" by the regress test runner. This affects the application name reported by pg_stat_replication when a pg_basebackup command is run from an isolation2 test (isolation2 uses pg_regress as test runner). Also, accept xlog_method as a parameter instead of hard-coding it to stream.
-
由 Asim R P 提交于
Comments in an isolation2 spec should either start from the end of a SQL command or from the beginning of a new line. Although this is only a subset of valid SQL comments, the change avoids a bug in parsing when a spec contains the string "--" that is not a comment. E.g. command line with a long option (e.g. --target-gp-dbid=100). Fixed a few tests. Their answer files recorded syntax errors due to incorrect comment identification prior to this commit.
-
- 25 5月, 2019 7 次提交
-
-
由 Ashuka Xue 提交于
-
由 Bhuvnesh Chaudhary 提交于
-
由 Jamie McAtamney 提交于
Check that the target passed in is valid by comparing it against secrets filenames. Co-authored-by: Jamie McAtamney <jmcatamney@pivotal.io Co-authored-by: Kalen Krempely <kkrempely@pivotal.io
-
由 Kalen Krempely 提交于
When branching gpdb simply update the gen_pipeline branch variable to freeze to the appropriate pipeline version.
-
由 Kalen Krempely 提交于
Rename pipeline_type to pipeline_target for clarity now that teams have their own concourse instance and the -t argument specifys the fly target. Also add two new flags: --pipeline_configuration, to make it easy for teams to make dev pipelines that match the prod pipeline, and --use_branch, to make it easier to have pipelines in flight for multiple branches without overlap. Co-authored-by: NJamie McAtamney <jmcatamney@pivotal.io> Co-authored-by: NShoaib Lari <slari@pivotal.io>
-
由 Kalen Krempely 提交于
Remove unneeded shell=True from subprocess command. This is safer to use as it does not rely on calling an unknown shell. Co-authored-by: NJamie McAtamney <jmcatamney@pivotal.io>
-
由 Kalen Krempely 提交于
Update gen_pipeline to take a team target for the pipeline_type parameter, which then uses the appropriate team's variable file. Co-authored-by: NJamie McAtamney <jmcatamney@pivotal.io>
-
- 24 5月, 2019 2 次提交
-
-
由 Adam Berlin 提交于
rsync is a dependency of the management utilities and was causing the gpinitmirror tests in isolation2 to fail.
-
由 Zhenghua Lyu 提交于
There are two reasons that we can remove this case to make pipeline a little faster: 1. We have lock modes related cases, if lockmodes case passed, it should imply the select-for-update case is correct 2. We have plan in future to enable isolation cases that test this.
-