external
/
citus
mirror of https://github.com/citusdata/citus.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
2,820 Commits
636 Branches
162 Tags
98 MiB
Commit Graph

1334 Commits (7fcbe92c597c6498a304a140dbf008011c8adcf7)

Author SHA1 Message Date
Marco Slot 7fcbe92c59 Disallow distributed functions with distribution arguments unless replication_model is streaming 2019-11-11 14:52:57 +01:00
Marco Slot f110a9fbe7 Do not try to sync metadata on standby coordinator 2019-11-11 14:52:25 +01:00
Jelte Fennema 1c7e02280f Include fmgr.h, don't duplicate FunctionCallInfo typedef 2019-11-11 14:51:16 +01:00
Hanefi Onaldi a853a824a5
Bump citus to 9.0.1 2019-10-25 10:36:18 +03:00
Marco Slot 2f201a98bb
Revoke usage from the citus schema 
(cherry picked from commit 04040e0a37)
 2019-10-25 10:06:09 +03:00
Onder Kalaci f8417d184c Fix memory leak on ReceiveResults 
It turns out that TupleDescGetAttInMetadata() allocates quite a lot
of memory. And, if the target list is long and there are too many rows
returning, the leak becomes appereant.

You can reproduce the issue wout the fix with the following commands:

```SQL

CREATE TABLE users_table (user_id int, time timestamp, value_1 int, value_2 int, value_3 float, value_4 bigint);
SELECT create_distributed_table('users_table', 'user_id');

insert into users_table SELECT i, now(), i, i, i, i FROM generate_series(0,99999)i;

-- load faster

-- 200,000
INSERT INTO users_table SELECT * FROM users_table;

-- 400,000
INSERT INTO users_table SELECT * FROM users_table;

-- 800,000
INSERT INTO users_table SELECT * FROM users_table;

-- 1,600,000
INSERT INTO users_table SELECT * FROM users_table;

-- 3,200,000
INSERT INTO users_table SELECT * FROM users_table;

-- 6,400,000
INSERT INTO users_table SELECT * FROM users_table;

-- 12,800,000
INSERT INTO users_table SELECT * FROM users_table;

-- making the target list entry wider speeds up the leak to show up
 select *,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,*,* FROM users_table ;

 ```
 2019-10-22 17:37:19 +02:00
Nils Dijk 4a4a220945
Fix enum add value order and pg12 (#3082) 
DESCRIPTION: Fix order for enum values and correctly support pg12

PG 12 introduces `ALTER TYPE ... ADD VALUE ...` during transactions. Earlier versions would error out when called in a transaction, hence we connect to workers outside of the transaction which could cause inconsistencies on pg12 now that postgres doesn't error with this syntax anymore.

During the implementation of this fix it became apparent there was an error with the ordering of enum labels when the type was recreated. A patch and test have been included.
 2019-10-07 17:16:19 +02:00
Jelte Fennema 01da11f264
Change citus truncate trigger to AFTER and add more upgrade tests (#3070) 
* Add more upgrade tests

* Fix citus trigger generation after upgrade

citus_truncate_trigger runs before truncate when created by create_distributed_table:
492d1b2cba/src/backend/distributed/commands/create_distributed_table.c (L1163)

* Remove pg_dist_jobid_seq
 2019-10-07 16:43:04 +02:00
Onder Kalaci 3be72ce42f Make sure that distributed functions always have the correct user 
Objectives:

(a) both super user and regular user should have the correct owner for the function on the worker
(b) The transactional semantics would work fine for both super user and regular user
(c) non-super-user and non-function owner would get a reasonable error message if tries to distribute the function

Co-authored-by: @serprex
 2019-10-04 21:38:49 +00:00
Marco Slot 1a3a174f67 Grant usage on schema citus to public 2019-10-04 12:26:08 +02:00
Marco Slot 89377ee578 Move RowExclusiveLock to start in SyncMetadataToNodes 2019-10-04 12:07:41 +02:00
Hadi Moshayedi 217db2a03e Don't block for locks in SyncMetadataToNodes() 2019-10-03 16:53:36 -07:00
Hadi Moshayedi ae915493e6 Don't send metadata commands to not-synced workers. 
Otherwise some of the dependencies might not exist yet and
commands will error out.
 2019-10-03 16:52:25 -07:00
Marco Slot 0b4b63e647 Drop the rebalancer before creating new UDFs 2019-10-03 16:08:58 +02:00
Marco Slot 2e50306cf8 Check command type in TryToDelegateFunctionCall 2019-10-03 15:37:15 +02:00
Hanefi Onaldi bd416ef68f Fix empty FROM clauses in PG12 2019-10-01 19:54:11 +00:00
Jelte Fennema ec4a165eec Improve isolation test block detection (#3055) 2019-10-01 14:10:15 +02:00
Jelte Fennema 40f785e6d8 Move citus_isolation_test_session_is_blocked to separate udf sql file 2019-10-01 14:10:15 +02:00
Philip Dubé 89d35e9692 Attempt to force custom plans for prepared statements when trying to delegate function calls 
We discern between PARAM_EXEC & PARAM_EXTERN:
d52eaa0948/src/include/nodes/primnodes.h (L211)
According to primnodes.h we should only run into PARAM_EXEC or PARAM_EXTERN
 2019-09-30 23:49:14 +00:00
Philip Dubé 29f1ea079b PG_VERSION_NUM > 110000 should be PG_VERSION_NUM >= 110000 
Also fix a > 12000 typo
 2019-09-30 23:37:43 +00:00
Hadi Moshayedi 5e97e5c98e Don't push down queries when in subqueries/ctes 2019-09-30 14:22:05 -07:00
Marco Slot 35bef0f3db Avoid caching connections from backends that servicei internal connections 2019-09-28 08:32:10 +02:00
Nils Dijk 01b26cf91a
Disallow distributed functions for functions depending on an extension (#3049) 
DESCRIPTION: Disallow distributed functions for functions depending on an extension

Functions depending on an extension cannot (yet) be distributed by citus. If we would allow this it would cause issues with our dependency following mechanism as we stop following objects depending on an extension.

By not allowing functions to be distributed when they depend on an extension as well as not allowing to make distributed functions depend on an extension we won't break the ability to add new nodes. Allowing functions depending on extensions to be distributed at the moment could cause problems in that area.
 2019-09-30 15:19:47 +02:00
Nils Dijk 473cbc0115
Propagate CREATE OR REPLACE FUNCTION to workers for distributed functions (#3043) 
DESCRIPTION: Propagate CREATE OR REPLACE FUNCTION

Distributed functions could be replaced, which should be propagated to the workers to keep the function in sync between all nodes.

Due to the complexity of deparsing the `CreateFunctionStmt` we actually produce the plan during the processing phase of our utilityhook. Since the changes have already been made in the catalog tables we can reuse `pg_get_functiondef` to get us the generated `CREATE OR REPLACE` sql.
 2019-09-30 12:41:17 +02:00
Jelte Fennema 82ec918b29
Add explain summary support (#3046) 
Fixes #2922 and also adds explain analyze regression tests
 2019-09-30 10:58:49 +02:00
Nils Dijk 9c2c50d875
Hookup function/procedure deparsing to our utility hook (#3041) 
DESCRIPTION: Propagate ALTER FUNCTION statements for distributed functions

Using the implemented deparser for function statements to propagate changes to both functions and procedures that are previously distributed.
 2019-09-27 22:06:49 +02:00
Philip Dubé 363409a0c2 Propagate REINDEX TABLE & REINDEX INDEX 2019-09-27 18:14:53 +00:00
Hanefi Onaldi 66b9f2e887 Deparsing and qualifiying for FUNCTION/PROCEDURE statements (#3014) 
This PR aims to add all the necessary logic to qualify and deparse all possible `{ALTER|DROP} .. {FUNCTION|PROCEDURE}` queries.

As Procedures are introduced in PG11, the code contains many PG version checks. I tried my best to make it easy to clean up once we drop PG10 support.


Here are some caveats:
- I assumed that the parse tree is a valid one. There are some queries that are not allowed, but still are parsed successfully by postgres planner. Such queries will result in errors in execution time. (e.g. `ALTER PROCEDURE p STRICT` -> `STRICT` action is valid for functions but not procedures. Postgres decides to parse them nevertheless.)
 2019-09-27 19:02:52 +02:00
Marco Slot 2868e02a3d Implement SELECT function call delegation. 
When a function is marked as colocated with a distributed table,
we try delegating queries of kind "SELECT func(...)" to workers.

We currently only support this simple form, and don't delegate
forms like "SELECT f1(...), f2(...)", "SELECT f1(...) FROM ...",
or function calls inside transactions.

As a side effect, we also fix the transactional semantics of DO blocks.
Previously we didn't consider a DO block a multi-statement transaction.
Now we do.

Co-authored-by: Marco Slot <marco@citusdata.com>
Co-authored-by: serprex <serprex@users.noreply.github.com>
Co-authored-by: pykello <hadi.moshayedi@microsoft.com>
 2019-09-27 09:13:25 -07:00
Jelte Fennema dab16be283
Set default threshold on get_rebalance_table_shards_plan to 0, like rebalance_table_shards (#3039) 
In this PR the default `threshold` of `rebalance_table_shards` was set to 0: https://github.com/citusdata/shard_rebalancer/pull/73
However, the default for get_rebalance_table_shards_plan was not updated. This
can cause the confusing situation where the actual steps run by
`rebalance_table_shards` are not the same as the ones returned by
`get_rebalance_table_shards_plan`.
 2019-09-27 17:21:36 +02:00
Marco Slot 32a11bdf6c Return early for common commands in the utility hook (#3031) 
We started copying parse trees by default further on in `multi_ProcessUtility`. That's not a problem for maintenance command, but might register for things like `PREPARE` and `EXECUTE`, which might happen thousands of times per second. Add a few common commands to the check at the start.
 2019-09-26 11:43:35 +02:00
Philip Dubé 4f60e3a149 Feedback 2019-09-24 17:31:09 +00:00
Marco Slot ca478defeb Deparse CALL statement instead of using original query string 2019-09-24 17:31:09 +00:00
Philip Dubé 90e1f1442a Annotated tests for multi_mx_call. 
Co-authored-by: pykello <hadi.moshayedi@microsoft.com>
 2019-09-24 17:31:09 +00:00
Marco Slot e269d990c9 Cast the distribution argument value when possible 2019-09-24 17:31:09 +00:00
Philip Dubé 432a8ef85b Hadi's feedback 
Co-authored-by: pykello <hadi.moshayedi@microsoft.com>
Co-authored-by: serprex <serprex@users.noreply.github.com>
 2019-09-24 17:31:09 +00:00
Philip Dubé bc1ad67eb5 Distribute CALL on distributed procedures to metadata workers 
Lots taken from https://github.com/citusdata/citus/pull/2829
 2019-09-24 17:31:09 +00:00
Onder Kalaci 18de78f386 Relax the colocation checks for distributed functions 
As long as the types can be coerced, it is safe to pushdown
functions.
 2019-09-24 16:31:08 +02:00
Marco Slot 42be8afd74 Swap pg_dist_node groupid and nodeid sequences 2019-09-24 12:03:44 +02:00
Hadi Moshayedi 48078a30e6 Fix wait_until_metadata_sync() for postgres 12. 
Postgres 12 now has an assertion that the calls to WaitLatchOrSocket
handle postmaster death.
 2019-09-23 14:15:35 -07:00
Philip Dubé 06faba91c0 Include ifdefs for pg12 API changes, update local_shard_executiuon test to avoid CTE inlining 2019-09-23 20:22:35 +00:00
Onder Kalaci d37745bfc7 Sync metadata to worker nodes after create_distributed_function 
Since the distributed functions are useful when the workers have
metadata, we automatically sync it.

Also, after master_add_node(). We do it lazily and let the deamon
sync it. That's mainly because the metadata syncing cannot be done
in transaction blocks, and we don't want to add lots of transactional
limitations to master_add_node() and create_distributed_function().
 2019-09-23 18:30:53 +02:00
Marco Slot 5f23b951c7 Support serial and smallserial when syncing metadata 2019-09-23 17:39:21 +02:00
Marco Slot e58d76c5f6 Fix assert failure in bare SELECT FROM reference table FOR UPDATE in MX 2019-09-23 17:00:09 +02:00
Marco Slot d85d77634d Handle anonymous composite types on the target list 2019-09-23 14:53:02 +02:00
Onder Kalaci d7e2968120 Add parameters to create_distributed_function() 
With this commit, we're changing the API for create_distributed_function()
such that users can provide the distribution argument and the colocation
information.
 2019-09-22 21:53:33 +02:00
Onder Kalaci e1fe8d60b4 Make sure that functions are also listed in SupportedDependencyByCitus 
We've recently merged two commits, db5d03931d
and eccba1d4c3, which actually operates
on the very similar places.

It turns out that we've an integration issue, where master_add_node()
fails to replicate the functions to newly added node.
 2019-09-20 11:02:50 +02:00
Hadi Moshayedi d24cefd055 Set active snapshot before SyncMetadataToNodes(). 2019-09-19 09:00:25 -07:00
Hanefi Onaldi ed11b9590c
Add distributed func creation queries in dependency replication logic 2019-09-18 20:07:45 +03:00
Hadi Moshayedi d2f2acc4b2 Make master_update_node citus-ha friendly. 2019-09-18 09:32:54 -07:00