Uh oh!
There was an error while loading. Please reload this page.
HBASE-27732 NPE in TestBasicWALEntryStreamFSHLog.testEOFExceptionInOl… - #5119
Conversation
Apache-HBase
commented
Mar 19, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 19, 2023
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 19, 2023
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 20, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 20, 2023
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 20, 2023
💔 -1 overall
This message was automatically generated. |
virajjasani
left a comment
There was a problem hiding this comment.
Looks like mvninstall failure above is flaky one right?
virajjasani
left a comment
There was a problem hiding this comment.
Changes look good overall
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
virajjasani
commented
Mar 21, 2023
I am +1 for the changes, but shall we update PR/Jira title to reflect changes done beyond test fix? Something like |
Apache9
commented
Mar 22, 2023
We could add more detailed information in the commit message as well as on the jira issue's description. Let me update. |
Apache9
commented
Mar 22, 2023
@virajjasani PTAL again. Thanks. |
Apache-HBase
commented
Mar 22, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 22, 2023
💔 -1 overall
This message was automatically generated. |
Apache9
commented
Mar 22, 2023
Let me check the failed UTs. |
Apache-HBase
commented
Mar 22, 2023
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 22, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 22, 2023
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 22, 2023
💔 -1 overall
This message was automatically generated. |
Apache9
commented
Mar 22, 2023
Oh, there is a data race if we also call cleanOldLogs in closeExecutor... Let me think how to better fix this, maybe a simple synchronized is enough as cleanOldLogs are all in memory operations after we introduced a special thread pool for archiving. |
Apache-HBase
commented
Mar 22, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 22, 2023
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 22, 2023
💔 -1 overall
This message was automatically generated. |
…dWALsDirectory Add a 'closed' flag in WALProps in AbstractFSWAL to indicate that whether a WAL file has been closed, if not, we will not try to archive it. Will mark it as closed after we fully close it in the background close task, and try to archive again. Also modified some tests since now the archiving of a rolled WAL file is also asynchronous, we need to wait instead of asserting directly.
Apache-HBase
commented
Mar 23, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 23, 2023
🎊 +1 overall
This message was automatically generated. |
virajjasani
commented
Mar 23, 2023
Okk, was this the only change |
Apache-HBase
commented
Mar 23, 2023
💔 -1 overall
This message was automatically generated. |
Apache9
commented
Mar 23, 2023
Yes, I added more comments for this method to explain why we need synchronized here. The failed UT is TestFuzzyRowFilterEndToEnd, which is not related. Let me trigger the UT again to confirm that there are no new flaky UTs. |
Apache-HBase
commented
Mar 23, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 23, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Mar 23, 2023
🎊 +1 overall
This message was automatically generated. |
virajjasani
commented
Mar 23, 2023
+1, thanks @Apache9 |
…dWALsDirectory (#5119) Add a 'closed' flag in WALProps in AbstractFSWAL to indicate that whether a WAL file has been closed, if not, we will not try to archive it. Will mark it as closed after we fully close it in the background close task, and try to archive again. Also modified some tests since now the archiving of a rolled WAL file is also asynchronous, we need to wait instead of asserting directly. Signed-off-by: Viraj Jasani <vjasani@apache.org> (cherry picked from commit 230fdc0)
…dWALsDirectory (#5119) Add a 'closed' flag in WALProps in AbstractFSWAL to indicate that whether a WAL file has been closed, if not, we will not try to archive it. Will mark it as closed after we fully close it in the background close task, and try to archive again. Also modified some tests since now the archiving of a rolled WAL file is also asynchronous, we need to wait instead of asserting directly. Signed-off-by: Viraj Jasani <vjasani@apache.org> (cherry picked from commit 230fdc0)
…dWALsDirectory (#5119) Add a 'closed' flag in WALProps in AbstractFSWAL to indicate that whether a WAL file has been closed, if not, we will not try to archive it. Will mark it as closed after we fully close it in the background close task, and try to archive again. Also modified some tests since now the archiving of a rolled WAL file is also asynchronous, we need to wait instead of asserting directly. Signed-off-by: Viraj Jasani <vjasani@apache.org> (cherry picked from commit 230fdc0)
…dWALsDirectory (apache#5119) Add a 'closed' flag in WALProps in AbstractFSWAL to indicate that whether a WAL file has been closed, if not, we will not try to archive it. Will mark it as closed after we fully close it in the background close task, and try to archive again. Also modified some tests since now the archiving of a rolled WAL file is also asynchronous, we need to wait instead of asserting directly. Signed-off-by: Viraj Jasani <vjasani@apache.org>
…dWALsDirectory (apache#5119) Add a 'closed' flag in WALProps in AbstractFSWAL to indicate that whether a WAL file has been closed, if not, we will not try to archive it. Will mark it as closed after we fully close it in the background close task, and try to archive again. Also modified some tests since now the archiving of a rolled WAL file is also asynchronous, we need to wait instead of asserting directly. Signed-off-by: Viraj Jasani <vjasani@apache.org> (cherry picked from commit 230fdc0) (cherry picked from commit e95b47e) Change-Id: I29133cfe4b4695bf817b7abb1cca775f4a5eb88a
…dWALsDirectory