a5370f5bb0
Realtime executor honours multi_shard_modify_mode
...
We're relying on multi_shard_modify_mode GUC for real-time SELECTs.
The name of the GUC is unfortunate, but, adding one more GUC
(or renaming the GUC) would make the UX even worse. Given that this
mode is mostly important for transaction blocks that involve modification
/DDL queries along with real-time SELECTs, we can live with the confusion.
2018-06-06 14:59:54 +03:00
d918556dca
INSERT .. SELECT pushdown honors multi_shard_modification_mode
2018-06-06 12:42:23 +03:00
336044f2a8
master_modify_multiple_shards() and TRUNCATE honors multi_shard_modification_mode
2018-06-06 12:29:05 +03:00
87270911d3
Merge pull request #2189 from citusdata/seq_ddl_via_seq_router
...
DDL commands honour `citus.multi_shard_modify_mode`
2018-06-06 09:25:26 +03:00
51cb24b39c
Increase timeout to make the appveyor tests happy
2018-06-05 17:52:18 +03:00
df44956dc3
Make sure that sequential DDL opens a single connection to each node
...
After this commit DDL commands honour `citus.multi_shard_modify_mode`.
We preferred using the code-path that executes single task router
queries (e.g., ExecuteSingleModifyTask()) in order not to invent
a new executor that is only applicable for DDL commands that require
sequential execution.
2018-06-05 17:52:17 +03:00
98b99634f3
Merge pull request #2167 from citusdata/serf
...
Add a debug message with distribution column value
2018-06-05 17:38:54 +03:00
fd4ff29f2f
Add a debug message with distribution column value
2018-06-05 15:09:17 +03:00
37bd4cd80c
Merge pull request #2195 from citusdata/push_grant
...
Add handling for grant/revoke all tables in schema
2018-05-31 14:26:52 +03:00
ba50e3f33e
Add handling for grant/revoke all tables in schema
2018-05-31 13:47:02 +03:00
b90c54abc0
Merge pull request #2170 from citusdata/master-update-version-1526383730
...
Bump Citus to 7.5devel
cr: @jasonmp85
2018-05-28 18:11:09 -06:00
49c85c2522
Fix windows citus_version
2018-05-28 17:43:39 -06:00
28e3003311
Update version in the root control
2018-05-28 17:25:21 -06:00
20acee2cd4
Bump citus version to 7.5devel
2018-05-28 17:25:21 -06:00
f069e7efb6
Merge pull request #2178 from citusdata/master-update-72
...
Add changelog entry for 7.2.2
2018-05-17 11:07:39 +03:00
673184b8d3
Add changelog entry for 7.2.2
2018-05-17 10:35:15 +03:00
cb634bf113
Merge pull request #2176 from citusdata/fix-appveyor-real-time-flakyness
...
Increase deadlock timeout so we get fewer signals
2018-05-17 09:49:11 +03:00
a7e09d777b
Increase deadlock timeout so we get fewer signals
2018-05-16 17:07:24 -07:00
9667ee5ac9
Alleviate OOM failures in COMMIT callback
...
Previously those failures caused us to crash, postgres abort()s when it
notices a failure in the COMMIT callback.
2018-05-15 16:39:33 -07:00
f163932a1b
Merge pull request #2169 from citusdata/citus-7.4.0-changelog-1526370810
...
Add changelog entry for 7.4.0
2018-05-15 14:18:07 +03:00
c248c8ac38
Add changelog entry for 7.4.0
2018-05-15 11:44:30 +03:00
cce658ad8c
Merge pull request #2166 from citusdata/version_bump_windows
...
Fix regression tests in AppVeyor
2018-05-11 22:27:48 +02:00
9323db2e05
Merge pull request #2161 from citusdata/move-intermediate-results-directory
...
Move call to RemoveIntermediateResultsDirectory
2018-05-11 11:44:53 +02:00
61d2c0f618
Stabilise output of multi_shard_update_delete test
2018-05-11 08:33:23 +02:00
ed47e4e6b9
Remove placementId from the ORDER BY to make results consistent
2018-05-11 17:04:50 +03:00
12e50d96dc
Fix tests where hll is not installed
2018-05-11 16:01:47 +03:00
b1619e182d
Make sure windows are build with the latest version
2018-05-11 15:44:19 +03:00
4c2bf5d2d6
Move call to RemoveIntermediateResultsDirectory
...
Errors thrown in the COMMIT handler will cause Postgres to segfault,
there's nothing it can do it abort the transaction by the time that
handler is called!
RemoveIntermediateResultsDirectory is problematic for two reasons:
- It has calls to ereport(ERROR which have been known to trigger
- It makes memory allocations which raise ERRORs when they fail
Once the COMMIT process has begun we don't use the intermediate results,
so it's safe to remove them a little earlier in the process. A failure
here will abort the transaction. That's pretty unnecessary, it's not
that important that we remove the results, but it's still better than a
crash.
2018-05-10 19:28:41 -07:00
b3e85e4f71
Fix the huge appveyor diffs
2018-05-10 18:18:43 -07:00
bdfe1ed702
Merge pull request #2165 from citusdata/enterprise_test_fix_drop_table_add
...
enterprise test fixes
2018-05-10 14:22:59 +03:00
b8c3197399
enterprise test fixes
2018-05-10 13:06:54 +03:00
0e1a01921d
Merge pull request #2164 from citusdata/fix_concurrent_reg_test
...
Run concurrent modification queries in tests sequentially
2018-05-10 12:17:41 +03:00
04d9e886fe
Run concurrent modification queries in tests sequentially
2018-05-10 11:59:18 +03:00
42b9690552
Merge pull request #2143 from citusdata/create_distributed_table_test_update_2
...
Create distributed table test update
2018-05-10 11:39:13 +03:00
785a86ed0a
Tests are updated to use create_distributed_table
2018-05-10 11:18:59 +03:00
ae97df43be
Merge pull request #2155 from citusdata/valgrind_fix
...
valgrind tests fixed
2018-05-10 10:46:02 +03:00
d35f2725bf
valgrind tests fix
2018-05-10 10:20:14 +03:00
a63e628120
Merge pull request #2142 from citusdata/master_update_node_locking
...
Make master_update_node block writes to the node
2018-05-09 14:27:03 +02:00
8b258cbdb0
Lock reads and writes only to the node being updated in master_update_node
...
Rather than locking out all the writes in the cluster, the function now only
locks out writes that target shards hosted by the node we're updating.
2018-05-09 15:14:20 +02:00
4198ad7618
Merge pull request #2157 from citusdata/fix_router_select
...
Throw an error if placements cannot be found in router executor
2018-05-08 23:04:45 -04:00
5f5f7b4fe0
Throw an error if placements cannot be found in router executor
2018-05-08 22:39:18 -04:00
b4cfa2f283
Merge pull request #2158 from citusdata/fix_mx_test
...
Run recursive_dml_queries_mx test on its own
2018-05-08 17:57:21 +02:00
b86d6eb544
Merge pull request #2151 from citusdata/fix_cte_xact
...
Ensure sigle-shard modifying CTEs are part of distributed transaction
2018-05-08 14:04:40 +02:00
5b079c125b
Merge pull request #2150 from citusdata/modify_volatility_check
...
Check volatile functions in modify queries
2018-05-08 11:45:36 +03:00
caa27161ca
Check volatile functions in modify queries
2018-05-08 11:16:40 +03:00
a7e6689890
Run recursive_dml_queries_mx test on its own
2018-05-06 17:12:53 +02:00
9438e5bde9
Ensure single-shard modifying CTEs are part of distributed transaction
2018-05-06 12:49:40 +02:00
86b12bc2d0
Always prefix operators with their namespace. ( #2147 )
...
Previously we checked if an operator is in pg_catalog, and if it wasn't we prefixed it with namespace in worker queries. This can have a huge impact on performance of physical planner when using custom data types.
This happened regardless of current search_path config, because Citus overrides the search path in get_query_def_extended(). When we do so, the check for existence of the operator in current search path in generate_operator_name() fails for any operators outside pg_catalog. This means that nothing gets cached, and in the following calls we will again recheck the system tables for existence of the operators, which took an additional 40-50ms for some of the usecases we were seeing.
In this change we skip the pg_catalog check, and always prefix the operator with its namespace.
2018-05-05 13:27:26 -04:00
0f98e4dd2f
Merge pull request #2137 from citusdata/marco_updel_subquery
...
Implement recursive planning for DML statements
2018-05-03 22:22:14 +02:00
42a8082721
PG11 compatibility refresh
...
adds a shim for a changed function api
2018-05-03 13:21:15 -06:00
Onder Kalaci
Murat Tuncer
Marco Slot
Murat Tuncer
Jason Petersen
Jason Petersen
velioglu
Burak Velioglu
Brian Cloutier
Marco Slot
Mehmet Furkan ŞAHİN
mehmet furkan şahin
Dimitri Fontaine
Hadi Moshayedi