[
https://issues.apache.org/jira/browse/OMID-139?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16806366#comment-16806366
]
Lars Hofhansl edited comment on OMID-139 at 4/1/19 3:52 AM:
------------------------------------------------------------
Cool!
I wasn't sure whether you needed the familyDeleteCache. (That seems to be a
per-row cache...?)
1000 is fine. Although it's always best to leave it to HBase and do it size
based. If you just write with autocommit on (on the HTable object) with a final
flush then HBase will pick the "best" flush size (at least you configure it
once for HBase)
was (Author: lhofhansl):
Cool!
I wasn't sure whether you needed the familyDeleteCache. (That seems to be a
per-row cache...?)
1000 is fine. Although it's always best to leave it to HBase and do it size
based. If you just write with autocommit on with a final flush then HBase will
pick the "best" flush size (at least you configure it once for HBase)
> Rollback is extremely slow
> --------------------------
>
> Key: OMID-139
> URL: https://issues.apache.org/jira/browse/OMID-139
> Project: Apache Omid
> Issue Type: Bug
> Affects Versions: 1.0.1
> Reporter: Lars Hofhansl
> Priority: Major
> Attachments: 138_139.patch
>
>
> Almost 3 minutes for about 100k rows.
> Haven't debugged it, yet, but it is possible that this is not batching the
> cleanup and doing individual RPCs?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
