roachtest: allocate new cluster when c.Extend fails#115539
roachtest: allocate new cluster when c.Extend fails#115539craig[bot] merged 1 commit intocockroachdb:masterfrom
Conversation
|
This change is |
|
Note that this change moves the cluster extend check from right before the start of the test (in Tested on local gceworker with error injected in c.Extend() and confirmed it works. Also running a 0.1 select probability test run in case the above note does break something. |
0da761b to
b627376
Compare
srosenberg
left a comment
There was a problem hiding this comment.
Reviewable status:
complete! 1 of 0 LGTMs obtained (waiting on @DarrylWong and @herkolategan)
pkg/cmd/roachtest/test_runner.go line 595 at r1 (raw file):
// after the test finishes so that the next test can be selected. If it // doesn't, extend it. timeout := testTimeout(&testToRun.spec)
It might improve readability slightly to factor this out into a separate helper.
pkg/cmd/roachtest/test_runner.go line 609 at r1 (raw file):
// We don't release the quota allocation - the new cluster will be // identical. testToRun.canReuseCluster = false
This block of code is duplicated with the above; it could be consolidated below in case either reuse prep. attempt (i.e.,Wipe or Extend) fails.
d33bc83 to
be018f4
Compare
DarrylWong
left a comment
There was a problem hiding this comment.
Reviewable status:
pkg/cmd/roachtest/test_runner.go line 595 at r1 (raw file):
Previously, srosenberg (Stan Rosenberg) wrote…
It might improve readability slightly to factor this out into a separate helper.
Good idea since WipeForReuse is already in a helper.
Done
pkg/cmd/roachtest/test_runner.go line 609 at r1 (raw file):
Previously, srosenberg (Stan Rosenberg) wrote…
This block of code is duplicated with the above; it could be consolidated below in case either reuse prep. attempt (i.e.,
WipeorExtend) fails.
Done
Previously, failure to extend the lifetime of a cluster would propagate to the worker and end up aborting the entire run. Now, we destroy the cluster and let it allocate a new one. Fixes: cockroachdb#112509 Release note: none Epic: none
be018f4 to
11f202c
Compare
|
Reconfirmed it works with the refactors. TFTR! bors r=srosenberg |
|
Build succeeded: |
115539: roachtest: allocate new cluster when c.Extend fails r=srosenberg a=DarrylWong Previously, failure to extend the lifetime of a cluster would propagate to the worker and end up aborting the entire run. Now, we destroy the cluster and let it allocate a new one. Fixes: cockroachdb#112509 Release note: none Epic: none 115649: cluster-ui: remove leading `/` in settings request r=xinhaoz a=xinhaoz Recently we required that all requests paths from cluster-ui do not start with `/` so that they may be used with any given subpath when constructing proxy requests. We missed removing the prefix from the `settings` api which requests all cluster settings. Release note: None Epic: none ---------------------- Previously, this auto stats enabled label would show disabled even though auto stats collection is enabled. This is because this value comes from the settings request which was not being issued correctly. https://www.loom.com/share/42847f2c63484a508a128c5808df7bdd 115654: opt: fix insert fast path uniqueness check bug r=mgartner a=mgartner This commit fixes a bug caused by creating a closure over an incrementing integer in a loop. Fixes cockroachdb#115377 Fixes cockroachdb#115378 Release note (bug fix): A bug has been fixed that caused node crashes and panics when running `INSERT` queries on `REGIONAL BY ROW` tables with `UNIQUE` constraints or indexes. The bug is only present in version 23.2.0-beta.1. Co-authored-by: DarrylWong <darryl@cockroachlabs.com> Co-authored-by: Xin Hao Zhang <xzhang@cockroachlabs.com> Co-authored-by: Marcus Gartner <marcus@cockroachlabs.com>
|
blathers backport 23.2 23.1 |
|
Encountered an error creating backports. Some common things that can go wrong:
You might need to create your backport manually using the backport tool. error creating merge commit from 11f202c to blathers/backport-release-23.1-115539: POST https://api.github.com/repos/cockroachdb/cockroach/merges: 409 Merge conflict [] you may need to manually resolve merge conflicts with the backport tool. Backport to branch 23.1 failed. See errors above. 🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf. |
Previously, failure to extend the lifetime of a cluster would propagate to the worker and end up aborting the entire run. Now, we destroy the cluster and let it allocate a new one.
Fixes: #112509
Release note: none
Epic: none