[jira] [Comment Edited] (PHOENIX-4219) Index gets out of sync on HBase 1.x

["James Taylor (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/PHOENIX-4219?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16182110#comment-16182110 ] 

James Taylor edited comment on PHOENIX-4219 at 9/27/17 4:24 PM:
----------------------------------------------------------------

[~samarthjain] - I've narrowed this down a bit and can only repro when column encoding is enabled:
- it works fine in all branches with column encoding disabled
- it works fine even with column encoding in 0.98
- it works fine if only a single thread is updating the table
- it works fine for immutable tables (with both local and global index).
- it's not a regression as the test fails on 4.11 as well.

Do you think you could take a look? I've attached a slightly modified version of the test because I was worried about potential race conditions if the client thread gets interrupted while the server is still writing index updates. This version prevents that.


was (Author: jamestaylor):
[~samarthjain] - I've narrowed this down a bit and can only repro when column encoding is enabled:
- it works fine in all branches with column encoding disabled
- it works fine even with column encoding in 0.98
- it works fine if only a single thread is updating the table
- it works fine for immutable tables (with both local and global index).

Do you think you could take a look? I've attached a slightly modified version of the test because I was worried about potential race conditions if the client thread gets interrupted while the server is still writing index updates. This version prevents that.

> Index gets out of sync on HBase 1.x
> -----------------------------------
>
>                 Key: PHOENIX-4219
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-4219
>             Project: Phoenix
>          Issue Type: Bug
>    Affects Versions: 4.12.0
>            Reporter: Vincent Poon
>            Priority: Blocker
>         Attachments: PHOENIX-4219_test.patch, PHOENIX-4219_test_v2.patch
>
>
> When writing batches in parallel with multiple background threads, it seems the index sometimes gets out of sync.  This only happens on the master and 4.x-HBase-1.2.
> The tests pass for 4.x-HBase-0.98
> See the attached test, which writes with 2 background threads with batch size of 100.



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