REGNAMESPACE function not found with optimizer rules disabled #98322
Assignees
Labels
branch-master
Failures and bugs on the master branch.
C-test-failure
Broken test (automatically or manually discovered).
O-roachtest
O-robot
Originated from a bot.
T-sql-queries
SQL Queries Team
Milestone
Comments
cockroach-teamcity
added
branch-master
|
Simplified test case: The REGNAMESPACE function doesn't exist in v22.2, but isn't a blocker since a normalization rule has to be disabled to hit the problem. |
msirek
removed
the
release-blocker
msirek
changed the title
|
This problem is occurring when the So, I think for internal SQLs, this rule should never be disabled. |
msirek
pushed a commit
to msirek/cockroach
that referenced
this issue
Mar 21, 2023
Casting an OID to REGNAMESPACE runs this SQL using the internal executor: ``` SELECT pg_namespace.oid, nspname FROM pg_catalog.pg_namespace WHERE oid = $1 ``` When the `GenerateConstrainedScans` rule is disabled, this returns no rows and so doesn't apply the cast. The same SQL, not run via an internal executor, but also with `GenerateConstrainedScans` disabled, behaves correctly. Disabling `GenerateConstrainedScans` all the time prevents the cluster from starting. Internal SQL is sensitive to disabled optimizer rewrite rules, and it could cause server instability to allow rules to ever be disabled for internal SQL. Also, the `testing_optimizer_disable_rule_probability` session setting is mainly meant to perturb query plans of foreground SQL in randomized tests, and enabling it for internal SQL could make these tests more noisy and create harder to debug issues. Therefore the fix is to set the `testing_optimizer_disable_rule_probability` setting to zero for internal SQL statements. Fixes cockroachdb#98322 Release note: None
msirek
pushed a commit
to msirek/cockroach
that referenced
this issue
Mar 22, 2023
Casting an OID to REGNAMESPACE runs this SQL using the internal executor: ``` SELECT pg_namespace.oid, nspname FROM pg_catalog.pg_namespace WHERE oid = $1 ``` When the `GenerateConstrainedScans` rule is disabled, this returns no rows and so doesn't apply the cast. The same SQL, not run via an internal executor, but also with `GenerateConstrainedScans` disabled, behaves correctly. Disabling `GenerateConstrainedScans` all the time prevents the cluster from starting. Internal SQL is sensitive to disabled optimizer rewrite rules, and it could cause server instability to allow rules to ever be disabled for internal SQL. Also, the `testing_optimizer_disable_rule_probability` session setting is mainly meant to perturb query plans of foreground SQL in randomized tests, and enabling it for internal SQL could make these tests more noisy and create harder to debug issues. Therefore the fix is to set the `testing_optimizer_disable_rule_probability` setting to zero for internal SQL statements. Fixes cockroachdb#98322 Release note: None
msirek
pushed a commit
to msirek/cockroach
that referenced
this issue
Mar 22, 2023
Casting an OID to REGNAMESPACE runs this SQL using the internal executor: ``` SELECT pg_namespace.oid, nspname FROM pg_catalog.pg_namespace WHERE oid = $1 ``` When the `GenerateConstrainedScans` rule is disabled, this returns no rows and so doesn't apply the cast. The same SQL, not run via an internal executor, but also with `GenerateConstrainedScans` disabled, behaves correctly. Disabling `GenerateConstrainedScans` all the time prevents the cluster from starting. Internal SQL is sensitive to disabled optimizer rewrite rules, and it could cause server instability to allow rules to ever be disabled for internal SQL. Also, the `testing_optimizer_disable_rule_probability` session setting is mainly meant to perturb query plans of foreground SQL in randomized tests, and enabling it for internal SQL could make these tests more noisy and create harder to debug issues. Therefore the fix is to set the `testing_optimizer_disable_rule_probability` setting to zero for internal SQL statements. Fixes cockroachdb#98322 Release note: None
craig bot
pushed a commit
that referenced
this issue
Apr 5, 2023
99168: randgen: disable generation of REGNAMESPACE type expressions r=msirek a=msirek Casting an OID to REGNAMESPACE runs this SQL using the internal executor: ``` SELECT pg_namespace.oid, nspname FROM pg_catalog.pg_namespace WHERE oid = $1 ``` When the `GenerateConstrainedScans` rule is disabled, this returns no rows and so doesn't apply the cast. The same SQL, not run via an internal executor, but also with `GenerateConstrainedScans` disabled, behaves correctly. Disabling `GenerateConstrainedScans` all the time prevents the cluster from starting. The reason a full scan of pg_namespace doesn't find the OID is because it only checks for schemas in the current database: https://github.com/cockroachdb/cockroach/blob/7b341ffa678e4a22416e2274351fd56f415f5421/pkg/sql/virtual_schema.go#L602 https://github.com/cockroachdb/cockroach/blob/d10c3dd42c3dc40cad82792a30ae47fd2a663f43/pkg/sql/pg_catalog.go#L2099-L2106 https://github.com/cockroachdb/cockroach/blob/a7e9c4a68b81436d1f9382518d4267f74cbdac94/pkg/sql/information_schema.go#L2295-L2296 Whereas with a constrained scan, all descriptors are searched: https://github.com/cockroachdb/cockroach/blob/af6a72a622ae05f3733b5db637403b3eaa9455f1/pkg/sql/catalog/descs/descriptor.go#L168-L175 The correct result is from the constrained scan, because we currently allow cross-database references. Once #55791 is fixed, both results should match. The fix alternatives are: 1. disallow disabling of the `GenerateConstrainedScans` rule for internal SQL 2. turn off generation of REGNAMESPACE expressions in randgen to avoid hitting this problem in tests. The 2nd fix alternative is implemented to avoid losing any of the rule-disabling test coverage provided by the `testing_optimizer_disable_rule_probability` setting. Fixes #98322 100652: cluster-ui: fix cached data invalidation on timescale change r=xinhaoz a=xinhaoz In a prior change, we moved the invalidation of cached data depending on the timescale to the local storage saga for CC. This was so invaldiation would occur after updating the cache. The local storage saga created for the time scale action was not hooked up to fire after the action, thus the data would not have been invalidated. This commit properly subscribes the saga to the update time scale action in CC. Epic: none Release note: None 100760: build: make sure toolchain has correct arguments for linking shared lib r=rail a=rickystewart `geos` has been failing to build without this change since #100313. This fixes it. Epic: none Release note: None Co-authored-by: Mark Sirek <[email protected]> Co-authored-by: Xin Hao Zhang <[email protected]> Co-authored-by: Ricky Stewart <[email protected]>
blathers-crl bot
pushed a commit
that referenced
this issue
Apr 5, 2023
Casting an OID to REGNAMESPACE runs this SQL using the internal executor: ``` SELECT pg_namespace.oid, nspname FROM pg_catalog.pg_namespace WHERE oid = $1 ``` When the `GenerateConstrainedScans` rule is disabled, this returns no rows and so doesn't apply the cast. The same SQL, not run via an internal executor, but also with `GenerateConstrainedScans` disabled, behaves correctly. Disabling `GenerateConstrainedScans` all the time prevents the cluster from starting. The reason a full scan of pg_namespace doesn't find the OID is because it only checks for schemas in the current database: https://github.com/cockroachdb/cockroach/blob/7b341ffa678e4a22416e2274351fd56f415f5421/pkg/sql/virtual_schema.go#L602 https://github.com/cockroachdb/cockroach/blob/d10c3dd42c3dc40cad82792a30ae47fd2a663f43/pkg/sql/pg_catalog.go#L2099-L2106 https://github.com/cockroachdb/cockroach/blob/a7e9c4a68b81436d1f9382518d4267f74cbdac94/pkg/sql/information_schema.go#L2295-L2296 Whereas with a constrained scan, all descriptors are searched: https://github.com/cockroachdb/cockroach/blob/af6a72a622ae05f3733b5db637403b3eaa9455f1/pkg/sql/catalog/descs/descriptor.go#L168-L175 The correct result is from the constrained scan, because we currently allow cross-database references. Once #55791 is fixed, both results should match. The fix alternatives are: 1. disallow disabling of the `GenerateConstrainedScans` rule for internal SQL 2. turn off generation of REGNAMESPACE expressions in randgen to avoid hitting this problem in tests. The 2nd fix alternative is implemented to avoid losing any of the rule-disabling test coverage provided by the `testing_optimizer_disable_rule_probability` setting. Fixes #98322 Release note: None
blathers-crl bot
pushed a commit
that referenced
this issue
Apr 5, 2023
Casting an OID to REGNAMESPACE runs this SQL using the internal executor: ``` SELECT pg_namespace.oid, nspname FROM pg_catalog.pg_namespace WHERE oid = $1 ``` When the `GenerateConstrainedScans` rule is disabled, this returns no rows and so doesn't apply the cast. The same SQL, not run via an internal executor, but also with `GenerateConstrainedScans` disabled, behaves correctly. Disabling `GenerateConstrainedScans` all the time prevents the cluster from starting. The reason a full scan of pg_namespace doesn't find the OID is because it only checks for schemas in the current database: https://github.com/cockroachdb/cockroach/blob/7b341ffa678e4a22416e2274351fd56f415f5421/pkg/sql/virtual_schema.go#L602 https://github.com/cockroachdb/cockroach/blob/d10c3dd42c3dc40cad82792a30ae47fd2a663f43/pkg/sql/pg_catalog.go#L2099-L2106 https://github.com/cockroachdb/cockroach/blob/a7e9c4a68b81436d1f9382518d4267f74cbdac94/pkg/sql/information_schema.go#L2295-L2296 Whereas with a constrained scan, all descriptors are searched: https://github.com/cockroachdb/cockroach/blob/af6a72a622ae05f3733b5db637403b3eaa9455f1/pkg/sql/catalog/descs/descriptor.go#L168-L175 The correct result is from the constrained scan, because we currently allow cross-database references. Once #55791 is fixed, both results should match. The fix alternatives are: 1. disallow disabling of the `GenerateConstrainedScans` rule for internal SQL 2. turn off generation of REGNAMESPACE expressions in randgen to avoid hitting this problem in tests. The 2nd fix alternative is implemented to avoid losing any of the rule-disabling test coverage provided by the `testing_optimizer_disable_rule_probability` setting. Fixes #98322 Release note: None
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
roachtest.unoptimized-query-oracle/disable-rules=half/rand-tables failed with artifacts on master @ 4dc9e98d8cc1b137ee4cc950a4a42e5895bb890b:
Parameters:
ROACHTEST_cloud=gce,ROACHTEST_cpu=4,ROACHTEST_encrypted=false,ROACHTEST_fs=ext4,ROACHTEST_localSSD=true,ROACHTEST_ssd=0Help
See: roachtest README
See: How To Investigate (internal)
This test on roachdash | Improve this report!
Jira issue: CRDB-25190
The text was updated successfully, but these errors were encountered: