-
Notifications
You must be signed in to change notification settings - Fork 3.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
kv/kvserver/protectedts/ptstorage: TestCorruptData failed #67972
Comments
@ajwerner we're getting two entries where one is expected by the test: cockroach/pkg/kv/kvserver/protectedts/ptstorage/storage_test.go Lines 576 to 583 in 10ede95
It seems silly - we're getting the "failed to parse row as record:" message twice, which, I don't know why there would even only ever be only one message? Don't we want to test that there is at least one message matching this pattern? Why are we expecting exactly one message to be logged? As far as I can tell this is operating on "the logger" that everyone and their dog can write to. The fix seems clear (what I said above) but want to check that I'm not missing anything. |
NB: we have facilities for more clever filtering nowadays. See for example what |
Yeah, I doubt there's any reason for it to be so strict. |
@AlexTalks this is another, more straightforward thing you could pick off. |
kv/kvserver/protectedts/ptstorage.TestCorruptData failed with artifacts on master @ 701b177d8f4b81d8654dfb4090a2cd3cf82e63a7:
ReproduceTo reproduce, try: make stressrace TESTS=TestCorruptData PKG=./pkg/kv/kvserver/protectedts/ptstorage TESTTIMEOUT=5m STRESSFLAGS='-timeout 5m' 2>&1 Parameters in this failure:
|
kv/kvserver/protectedts/ptstorage.TestCorruptData failed with artifacts on master @ 7897f24246bef3cb94f9f4bfaed474ecaa9fdee6:
ReproduceTo reproduce, try: make stressrace TESTS=TestCorruptData PKG=./pkg/kv/kvserver/protectedts/ptstorage TESTTIMEOUT=5m STRESSFLAGS='-timeout 5m' 2>&1 Parameters in this failure:
|
kv/kvserver/protectedts/ptstorage.TestCorruptData failed with artifacts on master @ ef22c9e3aaed735eb36f9a11da6af72396705e3e:
HelpSee also: [How To Investigate a Go Test Failure \(internal\)](https://cockroachlabs.atlassian.net/l/c/HgfXfJgM)Parameters in this failure:
|
kv/kvserver/protectedts/ptstorage.TestCorruptData failed with artifacts on master @ 46b7efa32c57b15f4ae6e0d75845f63360b62161:
HelpSee also: [How To Investigate a Go Test Failure \(internal\)](https://cockroachlabs.atlassian.net/l/c/HgfXfJgM)Parameters in this failure:
|
Sorry for ignoring this: #72503. |
72358: ccl/server/sql: Fix failing reset index usage unit test r=lindseyjin a=lindseyjin Resolves #72254 Previously, the unit test for resetting index usage statistics was failing. This was happening because of two reasons: 1. The IndexUsageStatsController was not being set properly for DistSQL server configs. 2. Noise from other tests was sometimes populating the index usage statistics table right after it was reset, causing the check for a newly reset table to sometimes randomly fail. To fix these issues, we have implemented the following: 1. Created a test case to reproduce the first error, and correctly populated IndexUsageStatsController for DistSQL server configs. 2. Updated the unit test to query for the specific test table id, to avoid other tables polluting the test space. The test has also been updated to check for a correct LastReset time. Release note: none 72466: opt: clean up FK ON UPDATE CASCADE code and opt trees r=mgartner a=mgartner #### opt: make FK ON UPDATE CASCADE input columns anonymous This commit anonymizes the columns in the input of a foreign key ON UPDATE CASCADE expression. This is safe because these columns can only be referenced by other expressions if they are update columns, and in that case, `mutationBuilder.addUpdateCascade` will give them a distinct name in the scope it produces. This change allows a special case added in #57153 to be removed, without failing the regression test. Release note: None #### opt: use more accurate FK ON UPDATE column metadata names Columns produced by the WithScan expression in a foreign key ON UPDATE CASCADE now have metadata names based on the column names of the child table, rather than the parent table. This more accurately describes the columns in the cascade plan because the plan is concerned with the child table, not the parent. Metadata names are only used in opt expression trees, so this is purely an aesthetic change, not a semantic one. Release note: None 72493: sessionphase: fix service latency computation in an error case r=yuzefovich a=yuzefovich If we encounter an error during the logical planning, then the end execution phase is never set. Previously, we would use that unset value in order to compute the latency of planning and execution and would get a negative result. This is now fixed. Release note: None 72503: protectedts/ptstorage: deflake test r=ajwerner a=ajwerner The test asserted that the error message appeared just once. If there's a retry the error can appear in the trace more than once. This commit cleans up that condition. Fixes #67972 Release note: None Co-authored-by: Lindsey Jin <[email protected]> Co-authored-by: Marcus Gartner <[email protected]> Co-authored-by: Yahor Yuzefovich <[email protected]> Co-authored-by: Andrew Werner <[email protected]>
kv/kvserver/protectedts/ptstorage.TestCorruptData failed with artifacts on master @ 2ca4acb22aade1175232187bb2b1754057e89f9b:
Reproduce
To reproduce, try:
Parameters in this failure:
This test on roachdash | Improve this report!
The text was updated successfully, but these errors were encountered: