roachtest: add MinVersion to scrub tests #33573
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The `scrub` tests have been passing on master, but failing on 2.1 because there was a bug (fixed in cockroachdb#32908). Release note: None
|
This change is |
vivekmenezes
approved these changes
Jan 9, 2019
complete! 1 of 0 LGTMs obtained
vivekmenezes
reviewed
Jan 9, 2019
|
bors r+ |
craig bot
pushed a commit
that referenced
this pull request
Jan 9, 2019
33573: roachtest: add MinVersion to scrub tests r=lucy-zhang a=lucy-zhang The `scrub` tests have been passing on master, but failing on 2.1 because there was a bug (fixed in #32908). Release note: None 33582: storage: bump RaftDelaySplitToSuppressSnapshotTicks r=knz a=knz The admin split path must accommodate a scenario where a range with not-yet-replicated followers is being manually split multiple times (eg. IMPORT during TPCC test fixtures). This scenario results in a bunch of replicas that all need to be populated with snapshots. To avoid backing up the raft snapshot queue, a heuristic was put in place (#32594) to delay the admin split if there is another snapshot being applied already. As shown by investigation in a failing test, there is a mismatch between the configured max delay for this heuristic (20s) and the actual duration of the snapshot application - the latter is limited by the max bandwidth for snapshots, 2 MB/s resulting in 32s applications in the worst case. We (Tobias and I) suspect that the heuristic thus fails to wait long enough to have the protective effect it was designed to provide. The current patch increases this delay to exceed this snapshot application duration estimate to about 50s. Note that this scenario is not likely to occur now that #32782 has been merged (this reduces the need for raft snapshots during splits); however in release-2.1 where that patch was not applied, the scenario is more likely. Release note (bug fix): resolve a cluster degradation scenario that could occur during IMPORT/RESTORE operations, manifested through a high number of pending Raft snapshots. Co-authored-by: Lucy Zhang <lucy-zhang@users.noreply.github.com> Co-authored-by: Raphael 'kena' Poss <knz@cockroachlabs.com>
Build succeeded |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The
scrubtests have been passing on master, but failing on 2.1 because therewas a bug (fixed in #32908).
Release note: None