importccl: IMPORT PGDUMP fails on ANALYZE #53419
Labels
A-disaster-recovery
A-spatial
Spatial work that is *not* related to builtins.
C-enhancement
Solution expected to add code/behavior + preserve backward-compat (pg compat issues are exception)
Comments
rytaft
added
C-enhancement
craig bot
pushed a commit
that referenced
this issue
Aug 25, 2020
52952: rowexec: joinReader optimizations for index joins r=sumeerbhola a=sumeerbhola These small optimizations are for index joins with large numbers of rows, which are common with geospatial queries. - No longer uses the keyToInputRowIndices map which was consuming about 1.5% in cpu profiles - Sorts the spans to use optimizations at the storage layer, like cockroachdb/pebble#860 - Avoids constructing the partial key string since it is not used Release note: None 53296: builtins: Implement ST_FlipCoordinates on arguments {geometry} r=otan a=themistoklik Should adopt PostGIS behavior. Release note (sql change): Implemented geometry builtin `ST_FlipCoordinates` Closes #48932 53307: implement ST_SharedPaths r=otan a=devenbhooshan Implemented the geometry based builtin `ST_SharedPaths` 53343: sql: remove auto-parallelism of UNION ALL r=yuzefovich a=yuzefovich Previously, `UNION ALL` construct was handled by simply merging the streams which could be handled by unordered synchronizers which are the source of concurrency in the engine. As a result, in some extreme cases (like when we have many UNION ALL constructs in the same query) we could have basically an unbounded parallelism that would stress the resources too much and could lead to OOM crashes. This issue is now fixed by using ordered synchronizers for the merging of the streams (which runs all inputs in the same goroutine). This approach is taken instead of implementing a separate UnionAll processor or operator because it is least invasive and should be easily backportable. Fixes: #51548. Release note (sql change): The concurrency of evaluation of UNION ALL queries has been reduced. Previously, such queries could end up crashing the server due to OOM in some extreme cases, and now that should be fixed at the expense of possible minor reduction in performance. 53358: sql: support the `USAGE` privilege on schemas r=rohany a=rohany Fixes #53342. This commit adds the `USAGE` privilege to schemas. The `USAGE` privilege allows a user to resolve objects under a schema. Release note (sql change): Support the `USAGE` privelege on schemas. 53421: importccl: ignore ANALYZE in PGDUMP r=mjibson a=mjibson Release note: None Fixes #53419 Co-authored-by: sumeerbhola <sumeer@cockroachlabs.com> Co-authored-by: Themis Papavasileiou <tpapavasileiou@bol.com> Co-authored-by: Deven Bhooshan <devenbhooshan@gmail.com> Co-authored-by: Yahor Yuzefovich <yahor@cockroachlabs.com> Co-authored-by: Rohan Yadav <rohany@alumni.cmu.edu> Co-authored-by: Matt Jibson <matt.jibson@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Due to #52823,
IMPORT PGDUMPis now failing whenever it sees anANALYZE table_namecommand. For example, I got the errorERROR: unsupported *redact.escapeArg statement: ANALYZE nyc_subway_stations. However, CockroachDB knows how to handleANALYZE table_name, so we should teach the import logic to allow it.cc @otan
The text was updated successfully, but these errors were encountered: