[jira] [Commented] (HBASE-19443) HBase Client gets stuck - Cannot get replica 0 location for...

["Lakshmiprasad D (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-19443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16333878#comment-16333878 ] 

Lakshmiprasad D commented on HBASE-19443:
-----------------------------------------

We have made a workaround to avoid this issue by starting a separate monitor thread. If the data loader thread hangs, monitor thread is interrupting the data loader thread and recreates a new data loader thread. Further the data loading goes ahead without any issues.

> HBase Client gets stuck - Cannot get replica 0 location for...
> --------------------------------------------------------------
>
>                 Key: HBASE-19443
>                 URL: https://issues.apache.org/jira/browse/HBASE-19443
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 1.2.4
>            Reporter: Lakshmiprasad D
>            Priority: Critical
>
> When the loader is loading data to HBase, without any forewarning that client gets stuck and prints the below message continuously:
> [ERROR] AsyncProcess[hconnection-0x71f5455a-shared--pool20-t40846] Cannot get replica 0 location for {"totalColumns":20,"families":{"v":[{"timestamp":1512584999999,"tag":[],"qualifier":"0","vlen":4},{"timestamp":1512584999999,"tag":[],"qualifier":"uid","vlen":8},{"timestamp":1512584999999,"tag":[],"qualifier":"1","vlen":8},{"timestamp":1512584999999,"tag":[],"qualifier":"2","vlen":11}]},"row":"C\\x99L[5\\x80\\x00\\x00"}
> At this time no other loaders are impacted. Restart of the loader seems to solve the issue until next occurence.
> Upon doing jstack at this time on the loader, it is observed that process is stuck on a object monitor inside 'AsyncProcess.java' class.



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