You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
150641: backup: remove useless fields in telemetry events r=kev-cao a=msbutler
Epic: none
Release note: none
151849: plpgsql: fix handling of annotations for DO blocks r=yuzefovich a=yuzefovich
**plpgsql: fix handling of annotations for DO blocks**
This commit should fix for good annotations handling for DO blocks that
was recently attempted to be fixed in 3b8ab38.
(That commit is partially reversed too).
The general issue is that in the PLpgSQL parser we use an ephemeral SQL
parser instance to process SQL statements. However, when processing the
DO block stmts we need to preserve the annotations state between each
stmt. This is now achieved by extending the SQL parser infrastructure to
give an option to override the initial annotations index as well as keep
on reusing the counter across multiple stmts. This allows us to
correctly count the number of annotations that we need for the whole DO
block during the initial parsing, and then we use the allocated
annotations container to set the items in the optbuild.
Fixes: #151848.
Release note (bug fix): Previously, CockroachDB node could crash when
executing DO stmts when they contain user-defined types (possibly
non-existing) in non-default configuration (additional logging like the
one controlled via `sql.log.all_statements.enabled` cluster setting needed
to be enabled). This bug was introduced in 25.1 release and is now fixed.
**tree: avoid index-of-bounds panics in GetAnnotation in production**
We've recently seen two panics when trying to access the Annotations
container and hitting the out-of-bounds condition. The annotations are
used to provide additional information about unresolved object names, so
if we simply return `nil` on an invalid access, we'll keep on using the
unresolved object name, which seems like a better option than process
crash. In test builds we'll keep on crashing.
Release note: None
153734: builtins: deflake TestGetSSTableMetricsSingleNode r=xinhaoz a=jbowens
Refactor and deflake TestGetSSTableMetricsSingleNode.
Fixes: #151742
Epic: none
Release note: none
153815: roachtest: delake pcr roachtests that offset reader tenant system tables r=jeffswenson a=msbutler
In this patch, we restart the reader tenant after the poller job bootstraps the reader catalog, to refresh the namespace cache. This is necessary because the reader tenant namespace cache believes the the priv table has id 100052 instead of the newly ingested id 52 from the source.
Fixes#153555
Epic: none
Release note: none
Co-authored-by: Michael Butler <[email protected]>
Co-authored-by: Yahor Yuzefovich <[email protected]>
Co-authored-by: Jackson Owens <[email protected]>
Copy file name to clipboardExpand all lines: docs/generated/eventlog.md
-5Lines changed: 0 additions & 5 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -3421,25 +3421,20 @@ logged whenever a BACKUP and RESTORE job completes or fails.
3421
3421
|`TargetScope`| TargetScope is the largest scope of the targets that the user is backing up or restoring based on the following order: table < schema < database < full cluster. | no |
3422
3422
|`IsMultiregionTarget`| IsMultiregionTarget is true if any of the targets contain objects with multi-region primitives. | no |
3423
3423
|`TargetCount`| TargetCount is the number of targets the in the BACKUP/RESTORE. | no |
3424
-
|`DestinationSubdirType`| DestinationSubdirType is - latest: if using the latest subdir - standard: if using a date-based subdir - custom: if using a custom subdir that's not date-based | no |
3425
3424
|`DestinationStorageTypes`| DestinationStorageTypes are the types of storage that the user is backing up to or restoring from. | no |
3426
3425
|`DestinationAuthTypes`| DestinationAuthTypes are the types of authentication methods that the user is using to access the destination storage. | no |
3427
3426
|`IsLocalityAware`| IsLocalityAware indicates if the BACKUP or RESTORE is locality aware. | no |
3428
-
|`AsOfInterval`| AsOfInterval is the time interval in nanoseconds between the statement timestamp and the timestamp resolved by the AS OF SYSTEM TIME expression. The interval is expressed in nanoseconds. | no |
3429
3427
|`WithRevisionHistory`| WithRevisionHistory is true if the BACKUP includes revision history. | no |
3430
3428
|`HasEncryptionPassphrase`| HasEncryptionPassphrase is true if the user provided an encryption passphrase to encrypt/decrypt their backup. | no |
3431
3429
|`KMSType`| KMSType is the type of KMS the user is using to encrypt/decrypt their backup. | no |
3432
3430
|`KMSCount`| KMSCount is the number of KMS the user is using. | no |
3433
3431
|`Options`| Options contain all the names of the options specified by the user in the BACKUP or RESTORE statement. For options that are accompanied by a value, only those with non-empty values will be present.<br><br>It's important to note that there are no option values anywhere in the event payload. Future changes to telemetry should refrain from adding values to the payload unless they are properly redacted. | no |
3434
-
|`DebugPauseOn`| DebugPauseOn is the type of event that the restore should pause on for debugging purposes. Currently only "error" is supported. | no |
3435
3432
|`JobID`| JobID is the ID of the BACKUP/RESTORE job. | no |
3436
3433
|`ResultStatus`| ResultStatus indicates whether the job succeeded or failed. | no |
3437
3434
|`ErrorText`| ErrorText is the text of the error that caused the job to fail. | partially |
3438
3435
|`RecurringCron`| RecurringCron is the crontab for the incremental backup. | no |
3439
3436
|`FullBackupCron`| FullBackupCron is the crontab for the full backup. | no |
3440
3437
|`CustomFirstRunTime`| CustomFirstRunTime is the timestamp for the user configured first run time. Expressed as nanoseconds since the Unix epoch. | no |
3441
-
|`OnExecutionFailure`| OnExecutionFailure describes the desired behavior if the schedule fails to execute. | no |
3442
-
|`OnPreviousRunning`| OnPreviousRunning describes the desired behavior if the previously scheduled BACKUP is still running. | no |
3443
3438
|`IgnoreExistingBackup`| IgnoreExistingBackup is true iff the BACKUP schedule should still be created even if a backup is already present in its destination. | no |
3444
3439
|`ApplicationName`| The application name for the session where recovery event was created. | no |
3445
3440
|`NumRows`| NumRows is the number of rows successfully imported, backed up or restored. | no |
0 commit comments