[
https://issues.apache.org/jira/browse/HBASE-8617?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13666930#comment-13666930
]
Ted Yu commented on HBASE-8617:
-------------------------------
{code}
+ if (isReplay) {
+ startRegionOperation();
+ } else {
+ startRegionOperation(Operation.BATCH_MUTATE);
{code}
There should be better name for BATCH_MUTATE enum: it reflects whether we're in
replay mode.
{code}
+
Assert.assertTrue(ioe.getMessage().contains("RegionInRecoveryException"));
{code}
Can we use (ioe instanceof RegionInRecoveryException) instead of checking
exception message ?
{code}
+ HTable ht = installTable(zkw, "table", "family", NUM_REGIONS_TO_CREATE);
{code}
I think zkw.close() should be called at the end of the test.
> Introducing a new config to disable writes during recovering
> -------------------------------------------------------------
>
> Key: HBASE-8617
> URL: https://issues.apache.org/jira/browse/HBASE-8617
> Project: HBase
> Issue Type: Improvement
> Components: regionserver
> Affects Versions: 0.98.0, 0.95.1
> Reporter: Jeffrey Zhong
> Assignee: Jeffrey Zhong
> Attachments: HBASE-8617.patch
>
>
> In distributedLogReplay(hbase-7006), we allow writes even when a region is in
> recovering. It may cause undesired behavior when applications(or deployments)
> already are near its write capacity because distributedLogReplay generates
> more write traffic to remaining region servers.
> The new config "hbase.regionserver.disallow.writes.when.recovering" tries to
> address the above situation so that recovering won't be affected by
> application normal write traffic.
> The default value of this config is false(meaning allow writes in recovery)
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
