hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Allan Yang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-18233) We shouldn't wait for readlock in doMiniBatchMutation in case of deadlock
Date Tue, 20 Jun 2017 02:47:00 GMT

    [ https://issues.apache.org/jira/browse/HBASE-18233?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16055073#comment-16055073
] 

Allan Yang commented on HBASE-18233:
------------------------------------

{quote}
Maybe return null directly in the else clause makes the logic clearer?
{quote}
Modified in the new patch
{quote}
Please also check the failed UT case and confirm whether relative. Thanks.
{quote}
The failed UTs are not related and passed locally. But the timeout case is related, I have
fixed this case in the new patch
{quote}
I'm not that sure but it seems we also have the issue in branch-1/branch-1.3?
{quote}
They may have the same issue, let me check, I think even in branch-2 and master branch have
this issue too. We should commit to this branch either.
[~carp84], thanks for your explanation, that's exactly what I’m thinking of!

{quote}
What about this question: "Where is the benefit? The benefit is that if some one comes through
looking for a write lock, first we'll flush any read lock batches? Any benefit seen?"
Is the benefit that we no longer block? That we make-do with however many locks we were able
to attain?
{quote}
Yes, you are right, boss [~stack]. As we discussed in HBASE-18144. The benefit is that we
won't create a block situation here. We should 'flush' batches with read locks and then come
back to try the failed readlock again. Even it is a readlock, blocking will create a temp
deadlock, right?
{quote}
Any benefit seen in testing?
{quote}
As you can see in the new added UT org.apache.hadoop.hbase.client.TestMultiParallel.testMultiThreadWithRowLocks,
if we wait for the lock, we will face 30 seconds timeout of getting locks
 


> We shouldn't wait for readlock in doMiniBatchMutation in case of deadlock
> -------------------------------------------------------------------------
>
>                 Key: HBASE-18233
>                 URL: https://issues.apache.org/jira/browse/HBASE-18233
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 1.2.7
>            Reporter: Allan Yang
>            Assignee: Allan Yang
>         Attachments: HBASE-18233-branch-1.2.patch, HBASE-18233-branch-1.2.v2.patch
>
>
> Please refer to the discuss in HBASE-18144
> https://issues.apache.org/jira/browse/HBASE-18144?focusedCommentId=16051701&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16051701



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message