loqrecovery: stage replica recovery plans on nodes #93044
Assignees
Labels
A-kv-replication
Relating to Raft, consensus, and coordination.
C-enhancement
Solution expected to add code/behavior + preserve backward-compat (pg compat issues are exception)
Comments
aliher1911
added
C-enhancement
|
cc @cockroachdb/replication |
craig bot
pushed a commit
that referenced
this issue
Jan 23, 2023
94390: loqrecovery: add storage for staged plans r=erikgrinaker a=aliher1911 Note: only the last commit belongs to this PR. This commit adds storage mechanism to persist replica update plans in the store directory to be able to apply upon node restart. Release note: None Part of #93044 94958: add CDC to add-to-project r=jlinder a=shermanCRL Issues & PRs will be added to the [new CDC project](https://github.com/orgs/cockroachdb/projects/38). GitHub Projects v2 has a different (non-REST) API and blathers doesn't do that. https://github.com/actions/add-to-project Epic: none 95675: backupccl: handle placeholders for revision_history r=benbardin,adityamaru a=stevendanna With the addition of ALTER BACKUP SCHEDULE, the revision_history option was changed to take an expression. However, we didn't update a few locations where we need to evaluate that expression. As a result, you could create a scheduled backup using a placeholder for the value of revision_history, but that scheduled backup would always fail because the placeholder value isn't available when the job is created. Epic: none Release note: Fix a bug in which CREATE SCHEDULE would not properly handle a placeholder for the revision_history option. Co-authored-by: Oleg Afanasyev <[email protected]> Co-authored-by: Matt Sherman <[email protected]> Co-authored-by: Steven Danna <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add plan stading as described in RFC.
Add Admin RPC to stage plan on all cluster nodes. Staging includes discovering all cluster nodes, verifying that there's no conflicting plan already staged and sending plan to every nodes. Every nodes then saves plan locally if there's any data in the plan eligible for application on that node, it also updated decommissioned status for nodes removed in local tombstone storage.
Extend apply-plan command to distribute plan in half-online mode.
Part of #77273
Jira issue: CRDB-22132
Epic CRDB-14205
The text was updated successfully, but these errors were encountered: