You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hbase.apache.org by "Ted Yu (JIRA)" <ji...@apache.org> on 2018/02/28 18:54:00 UTC

[jira] [Commented] (HBASE-20107) Add a test case for HBASE-14317

    [ https://issues.apache.org/jira/browse/HBASE-20107?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16380826#comment-16380826 ] 

Ted Yu commented on HBASE-20107:
--------------------------------

bq. `attainSafePoint` get out stuck.

Did you mean attainSafePoint would not get stuck ?

> Add a test case for HBASE-14317
> -------------------------------
>
>                 Key: HBASE-20107
>                 URL: https://issues.apache.org/jira/browse/HBASE-20107
>             Project: HBase
>          Issue Type: Improvement
>          Components: wal
>    Affects Versions: 3.0.0
>            Reporter: Zephyr Guo
>            Priority: Minor
>             Fix For: 3.0.0
>
>
> RS might get stuck because of  HBASE-14317. I met same problem in my hbase, and HBASE-14317 can solve this problmen very well.
> But I don't get stuck when I run `TestWalLockup.testLockupWhenSyncInMiddleOfZigZagSetup`.
> We'll never see stuck in this test case, because `FailedSyncBeforeLogCloseException` will break `zigzagLatch.waitSafePoint(SyncFuture)`. Finally, `zigzagLatch.releaseSafePoint()` will be call, and `attainSafePoint` get out stuck.
>  Here, we add a new test case to simulate stuck RS correctly.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)