HDFS-15654/15674: Fix TestBPOfferService#testMissBlocksWhenReregister#5089
HDFS-15654/15674: Fix TestBPOfferService#testMissBlocksWhenReregister#5089xinglin wants to merge 4 commits intoapache:branch-3.3from
Conversation
…ermittently (apache#2419) (cherry picked from commit 3248791)
…trunk. (apache#2467) (cherry picked from commit 425996e)
|
💔 -1 overall
This message was automatically generated. |
|
This PR only changed TestBPOfferService. It does not change any other files. The failed hdfs unit tests are probably due to flaky minihdfs. |
|
@xinglin - Can you trigger to trigger once again. Let's have a happy Yetus |
|
Hi @ashutoshcipher,
Made an empty commit to trigger another build. Let's whether we will get a happy yetus. |
|
💔 -1 overall
This message was automatically generated. |
|
It looks like four different unit tests failed the second time and there is no overlap between the first and second build. I guess unit tests were run in different orders between different builds and thus we see different unit tests failed each time. @ashutoshcipher, I don't think these non-deterministic unit tests failures are related with this PR, since we only modified TestBPOfferService.java. What do you think? |
|
Thanks @xinglin. Can you verify the failed tests once by running them in your local with the patch? |
TestDecommissionWithStripedBackoffMonitor is still failing sometimes. It is also reported in HDFS-15840 for trunk (unresolved). TestDFSInotifyEventInputStreamKerberized sometimes also failed. It is reported in HDFS-16167 for trunk (unresolved). TestExternalStoragePolicySatisfier sometimes failed due to timeout. But also got some success runs. If needed, we could increase the timeout for |
|
💔 -1 overall
This message was automatically generated. |
|
Tried a couple of times to run these unit tests locally. They did fail sometimes but also managed to get a few success runs. I don't think we should block this PR by these non-deterministic/not relevant unit tests failures. |
|
close this PR as two of these commits have been backported to branch-3.3 by others. |
Description of PR
Cherry-pick a couple of PRs from trunk to branch-3.3 for fixing TestBPOfferService#testMissBlocksWhenReregister
HDFS-15654: minor conflict due to
testCommandProcessingThreadExitbeing backported to branch-3.3 before HDFS-15654.HDFS-15674: clean cherry-pick.
HDFS-16224: clean cherry-pick.
How was this patch tested?
Run tests three times without error.
mvn test -Dtest="TestBPOfferService"