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
153065: restore: scale number of OR link workers to number of nodes r=msbutler a=kev-cao
This patch teaches online restore to scale the number of link workers to the number of nodes in the cluster instead of statically setting it to 32. In the new behavior, the default value for `backup.restore.online_worker_count` is 0, which sets the number of link workers to two times the number of nodes.
Epic: CRDB-48786Fixes: #146584
153380: sql: fix node panic on SIGINT during CHECK EXTERNAL CONNECTION r=jeffswenson a=kev-cao
`CHECK EXTERNAL CONNECTION` defers a close of the `rows` channel as part of its execution. It also closes that same channel in its `Close` method. Under normal execution, if `CHECK EXTERNAL CONNECTION` is allowed to complete, that `rows` channel is set to `nil`, so `Close` skips the attempt to close to the channel.
However, if a `SIGINT` is sent during the execution to cancel the query, `rows` is never set to `nil` and `Close` will attempt to close a closed channel, causing a node panic.
Epic: None
Release note (bug fix): Fixed a bug introduced in v25.1.0 that would cause a node panic if a `SIGINT` signal was sent during the execution of a `CHECK EXTERNAL CONNECTION` command.
153470: backupdest: attempt to close all stores in MakeBackupDestinationStores cleanupFn r=kev-cao a=msbutler
Epic: none
Release note: none
Co-authored-by: Kevin Cao <[email protected]>
Co-authored-by: Michael Butler <[email protected]>
0 commit comments