- 12 1月, 2018 3 次提交
-
-
由 Shreedhar Hardikar 提交于
This commit brings in ORCA changes that ensure that a Materialize node is not added under a Filter when its child contains outer references. Otherwise, the subplan is not rescanned (because it is under a Material), producing wrong results. A rescan is necessary for it evaluates the subplan for each of the outer referenced values. For example: ``` SELECT * FROM A,B WHERE EXISTS ( SELECT * FROM E WHERE E.j = A.j and B.i NOT IN ( SELECT E.i FROM E WHERE E.i != 10)); ``` For the above query ORCA produces a plan with two nested subplans: ``` Result Filter: (SubPlan 2) -> Gather Motion 3:1 -> Nested Loop Join Filter: true -> Broadcast Motion 3:3 -> Table Scan on a -> Table Scan on b SubPlan 2 -> Result Filter: public.c.j = $0 -> Materialize -> Result Filter: (SubPlan 1) -> Materialize -> Gather Motion 3:1 -> Table Scan on c SubPlan 1 -> Materialize -> Gather Motion 3:1 -> Table Scan on c Filter: i <> 10 ``` The Materialize node (on top of Filter with Subplan 1) has cdb_strict = true. The cdb_strict semantics dictate that when the Materialize is rescanned, instead of destroying its tuplestore, it resets the accessor pointer to the beginning and the subtree is NOT rescanned. So the entries from the first scan are returned for all future calls; i.e. the results depend on the first row output by the cross join. This causes wrong and non-deterministic results. Also, this commit reinstates this test in qp_correlated_query.sql. It also fixes another wrong result caused by the same issue. Note that the changes in rangefuncs_optimizer.out are because ORCA now no longer falls back for those queries. Instead it produces a plan which is executed on master (instead of the segments as was done by planner) which changes the error messages. Also bump ORCA version to 2.53.8. Signed-off-by: NEkta Khanna <ekhanna@pivotal.io>
-
由 Shoaib Lari 提交于
gpstart did a cluster-wide check of heap_checksum settings and refused to start the cluster if this setting was inconsistent. This meant a round of ssh'ing across the cluster which was causing OOM errors with large clusters. This commit moves the heap_checksum validation to gpsegstart.py, and changes the logic so that only those segments which have the same heap_checksum setting as master are started. Author: Nadeem Ghani <nghani@pivotal.io> Author: Shoaib Lari <slari@pivotal.io>
-
由 dyozie 提交于
-
- 11 1月, 2018 7 次提交
-
-
由 Larry Hamel 提交于
This reverts commit c4062620.
-
由 Goutam Tadi 提交于
- In the script used for creating user "gpadmin", instead of using a nonstandard group named "supergroup" that doesn't seem to be used anywhere, do the more standard practice of creating a group "gpadmin" that matches the username. Signed-off-by: NLarry Hamel <lhamel@pivotal.io>
-
由 Mel Kiyama 提交于
* docs: updates to create function attribute execute on * docs: review comment updates for CREATE FUNCTION
-
由 Jimmy Yih 提交于
Some of our OS container images have been updated to have conan 1.0.0. This new version changes the conan install command a bit.
-
由 Lisa Owen 提交于
-
由 Lisa Owen 提交于
-
由 Lisa Owen 提交于
* docs - type SET DEFAULT ENCODING * edits requested by david * add gp-specific clauses to summary of greenplum features page
-
- 10 1月, 2018 6 次提交
-
-
由 Heikki Linnakangas 提交于
It didn't handle relation forks correctly, and counted the main fork's size three times.
-
由 Heikki Linnakangas 提交于
-
由 Heikki Linnakangas 提交于
Now that we use the upstream implementation for window functions, the 'gp_enable_sequential_window_plans' and 'gp_idf_deduplicate' GUCs are no longer.
-
由 Lisa Owen 提交于
-
由 Dhanashree Kashid 提交于
HASHJOIN_IS_OUTER already existed in nodeHashJoin.c; but was added again by commit 321c0529. This commit removes the redundant copy of the same.
-
由 Jimmy Yih 提交于
This is to avoid collision of the nonsegwalrep binary.
-
- 09 1月, 2018 4 次提交
-
-
由 Shoaib Lari 提交于
Long ago, we thought we might need to support multiple mirrors. But we don't, and don't forsee it coming soon. Simplify the code to only ever have one mirror, but still allow for the possibility of no mirrors Author: Shoaib Lari <slari@pivotal.io> Author: C.J. Jameson <cjameson@pivotal.io>
-
由 Marbin Tan 提交于
The gpArray use of GpDB and Segment classes was confusing. This change renames GpDB to Segment and Segment to SegmentPair to clarify usage. Its a big diff, but a simple, repeating change. Author: Shoaib Lari <slari@pivotal.io> Author: Marbin Tan <mtan@pivotal.io> Author: C.J. Jameson <cjameson@pivotal.io>
-
由 Lav Jain 提交于
* Have a separate docker file for gpadmin user * Add indent package for centos6
-
由 Jacob Champion 提交于
The temporary file hack to deal with pygresql output oddities didn't ensure that those files were always removed. Instead of a named file we don't need, just use a TemporaryFile that will go away as soon as we release it. Added a FIXME to revisit this once we upgrade pygresql.
-
- 08 1月, 2018 1 次提交
-
-
由 Heikki Linnakangas 提交于
pg_proc_gp.h is generated from pg_proc.sql, but a few recent commits updated only pg_proc_gp.h, not pg_proc.sql. As a result, if you ran the perl script, the gp_replication_error() function vanished, and the obsolete PT verification functions reappeared. Update pg_proc_gp.h, so that pg_proc_gp.h is reproduced by the perl script again.
-
- 06 1月, 2018 5 次提交
-
-
由 Sambitesh Dash 提交于
Instead of assuming that casts are always binary coercible (and hence that we could get away with just dropping them), translate casts in ORCA plans into either a RelabelType or a FuncExpr. Signed-off-by: NSambitesh Dash <sdash@pivotal.io>
-
由 Mel Kiyama 提交于
PR for 5X_STABLE Will be ported to MAIN
-
由 Lisa Owen 提交于
-
由 Mel Kiyama 提交于
* docs: change GUC default: optimizer_join_arity_for_associativity_commutativity=18 * docs: GUC removed ZSTD (zstandard) from gp_default_storage_options (added by mistake)
-
由 Todd Sedano 提交于
The correct order is as follows: 1. modify ~/.bash_profile 2. ssh to localhost 3. make create-demo-cluster See https://github.com/greenplum-db/gpdb/issues/3704 for more detail
-
- 05 1月, 2018 7 次提交
-
-
由 Pengzhou Tang 提交于
There is a long-standing race condition in BackoffSweeper() which triggers an error and then triggers another assertion failure for not reset sweeperInProgress to false. This commit doesn't resolve the race condition fundamentally with lock or other implementation, because the whole backoff mechanism did not ask for accurate control, so skipping some sweeps should be fine so far. We also downgrade the log level to DEBUG because a restart of sweeper backend is unnecessary.
-
由 Kavinder Dhaliwal 提交于
-
由 Jesse Zhang 提交于
Partition tables hard-code the operator '=' lookup to namespace 'pg_catalog', which means that in this test we had to put our user-defined operator into that special system namespace. This works fine, until we try to pg_dump the resulting database: pg_catalog is not dumped by default. That led to an incomplete dump that will fail to restore. This commit reinstates the dropping of the schema at the end of `gporca` test to get the pipelines back to green (broken after c7ab6924 ). Backpatch to 5X_STABLE.
-
由 Daniel Gustafsson 提交于
xlogviewer is disconnected from the build, and hasn't built since we merged the HOT patch during the PostgreSQL 8.3 merge in the 5.x development cycle. Since we have contrib/xlogdump, remove rather than trying to resurrect it. Discussion: https://groups.google.com/a/greenplum.org/forum/#!topic/gpdb-dev/ov5Sxrd1JL4
-
由 Divya Bhargov 提交于
Since the docs folder is ignored, it is possible for the gpdb_src to get to a detatched HEAD state if there was a docs commit after a given commit. We don't really need the branch to get checkout the commit in question. Signed-off-by: NEd Espino <edespino@pivotal.io> Signed-off-by: NDivya Bhargov <dbhargov@pivotal.io>
-
由 Jesse Zhang 提交于
The `gporca` regression test suite uses a schema but doesn't really switch `search_path` to the schema that's meant to encapsulate most of the objects it uses. This has led to multiple instances where we: 1. Either used a table from another namespace by accident; 2. Or we leaked objects into the public namespace that other tests in turned accidentally depended on. As we were about to add a few user-defined types and casts to the test suite, we want to (at last) ensure that all future additions are scoped to the namespace. Signed-off-by: NSambitesh Dash <sdash@pivotal.io> Closes #4238
-
由 C.J. Jameson 提交于
Author: C.J. Jameson <cjameson@pivotal.io>
-
- 04 1月, 2018 5 次提交
-
-
由 Mel Kiyama 提交于
* docs: pl/container updates - plcontainer command, configuration file -removed experimental warning. -replaced plcontainer command with new options. -updated configuration file - changed element names, added setting element. -updated default shared volume. PR for 5X_STABLE Will be ported to MAIN * docs: pl/container - updates based on review comments. * docs: pl/container - more updates based on additional review comments. * docs: pl/container - minor edit * docs: pl/container - another minor fix.
-
由 Lisa Owen 提交于
* docs - add info about greenplum database release versioning * misc edits requested by david * update subnav titles * clarify that deprecated features removed only in major release
-
由 Mel Kiyama 提交于
* docs: gpstop new option --host * docs: gpstop - update update/clarify --host description based on review comments. * docs: gpstop --host. updates based on review comments. * docs: gpstop - added information on restoring segments after using --host. * docs: gpstop --host. corrected name of utility to recover segments : gprecoverseg.
-
由 Jacob Champion 提交于
Several \d variants don't put a row count at the end of their tables, which means that atmsort doesn't stop sorting output until it finds a row count somewhere later. Some tests are having their diff output horribly mangled because of this, which makes debugging difficult. When we see a \d command, try to apply more heuristics for finding the end of a table. In addition to ending at a row count, end table sorting whenever we find a line that doesn't have the same number of column separators as the table header. If we don't have a table header, still attempt to end table sorting at a blank line. extprotocol's test output order must be fixed as a result. Put the "External options" line where psql actually prints it, after "Format options".
-
由 David Sharp 提交于
https://github.com/hyperic/sigar is not under active development. https://github.com/boundary/sigar is a fork that has been somewhat updated. In particular, it has a fix to allow it to compile with GCC 5 and 6.
-
- 03 1月, 2018 2 次提交
-
-
由 Shreedhar Hardikar 提交于
* Fix 4 (out of 64) windowerr tests that use row_number() to be non-deterministic * Fix remaining 58 (out of 64) tests. There was a difference in the results between planner and optimizer due to different value of row_number assigned. row_number() is inherently non-deterministic in GPDB. For example, for the following query: select row_number() over (partition by b) from foo; Let's say that foo was not distributed by b. In this case, to compute the WindowAgg, we would first have to redistribute the table on b (or gather all the tuples on master). Thus, for rows having the same b value, the row_number assigned depends on the order in which they are received by the WindowAgg - which is non-deterministic. In qp_olap_windowerr.sql tests, we mitigate this by forcing an order on ord column, which is unique in this context, making it easier to compare test results. * Remove FIXME comment and enable optimizer_trace_fallback Signed-off-by: NShreedhar Hardikar <shardikar@pivotal.io>
-
由 Ed Espino 提交于
o This moves the validation (subprocess call) of the pipelines release jobs into the tool which generates the production pipeline. The corresponding task validate_pipeline.yml is removed and corresponding job. o Output production fly commands for both "gpdb_master" and "gpdb_master_without_asserts" pipelines. This will help engineers update both production pipelines. o Fix the icw sles jobs in the development pipelines. o Update README.md with usage examples. o Remove validate_pipeline from pr_pipeline.yml as this validation is moving to gen_pipeline.py.
-