[
https://issues.apache.org/jira/browse/PHOENIX-1402?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14196648#comment-14196648
]
Andrew Purtell edited comment on PHOENIX-1402 at 11/4/14 7:34 PM:
------------------------------------------------------------------
I've been wondering while chasing this down how we could have 2 or 4 minute
split transactions, they are supposed to happen quickly. See
https://issues.apache.org/jira/secure/attachment/12679262/HowHBaseRegionSplitsareImplemented.pdf
. From steps 4 through 9 the parent region is offline and the daughters are
not online yet. Any client trying to access the region during this period will
see NotServingRegionExceptions and will have to wait. Any significant
lengthening of the time the region is offline during a split transaction risks
client timeouts.
Edit: s/table/region/
was (Author: apurtell):
I've been wondering while chasing this down how we could have 2 or 4 minute
split transactions, they are supposed to happen quickly. See
https://issues.apache.org/jira/secure/attachment/12679262/HowHBaseRegionSplitsareImplemented.pdf
. From steps 4 through 9 the parent region is offline and the daughters are
not online yet. Any client trying to access the table during this period will
see NotServingRegionExceptions and will have to wait. Any significant
lengthening of the time the table is offline during a split transaction risks
client timeouts.
> Don't recalculate stats on split
> --------------------------------
>
> Key: PHOENIX-1402
> URL: https://issues.apache.org/jira/browse/PHOENIX-1402
> Project: Phoenix
> Issue Type: Bug
> Reporter: James Taylor
>
> Rather than scan the new regions on a split (which is potentially expensive,
> and might be causing the timeouts you're seeing [~jfernando_sfdc]), we should
> instead just split up the existing guideposts between new two new regions
> based on the split point.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
