[
https://issues.apache.org/jira/browse/PHOENIX-3706?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15898057#comment-15898057
]
Samarth Jain commented on PHOENIX-3706:
---------------------------------------
The reason we are seeing this error is because in 0.98, when we hit the hbase
rpc timeout, the hbase server side eventually ends up throwing an unknown
scanner exception which is a bug. Since 0.98 has EOL'ed, the only alternative
here is to set the hbase rpc timeout to a large value. Either way, this patch
addresses the issue of renew lease threads being stuck at the renewLease call
because the scanner has acquired the lock when calling scanner.next(). The
patch doesn't address the UnknownScannerException case. Will track that work in
another JIRA.
> RenewLeaseTask should give up and reattempt later to renewlease if lock
> cannot be acquired
> ------------------------------------------------------------------------------------------
>
> Key: PHOENIX-3706
> URL: https://issues.apache.org/jira/browse/PHOENIX-3706
> Project: Phoenix
> Issue Type: Bug
> Reporter: Samarth Jain
> Assignee: Samarth Jain
> Attachments: PHOENIX-3706.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
