You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Xu Cang (JIRA)" <ji...@apache.org> on 2018/09/24 08:47:00 UTC
[jira] [Commented] (HBASE-21193) Retrying Callable doesn't take max
retries from current context; uses defaults instead
[ https://issues.apache.org/jira/browse/HBASE-21193?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16625534#comment-16625534 ]
Xu Cang commented on HBASE-21193:
---------------------------------
[~stack]
tried a bit I think it's working for me unless I misunderstand what you mean.
I added a logline here for debugging:
[https://github.com/apache/hbase/blob/master/hbase-client/src/main/java/org/apache/hadoop/hbase/client/RpcRetryingCallerImpl.java#L79]
LOG.info("maxAttemps is : " + maxAttempts);
Then I modified hbas-site.config with as below
37 <property>
38 <name>hbase.client.retries.number</name>
39 <value>*12*</value>
40 </property>
Got this in log
276913 2018-09-24 01:39:56,012 INFO [master/192.168.0.9:16000:becomeActiveMaster] client.RpcRetryingCallerImpl: *maxAttemps is : 37*
Then change config to this
37 <property>
38 <name>hbase.client.retries.number</name>
*39 <value>5</value>*
40 </property>
Got maxAttemps changed as below:
279046 2018-09-24 01:41:55,552 INFO [master/192.168.0.9:16000:becomeActiveMaster] client.RpcRetryingCallerImpl: *maxAttemps is : 16*
> Retrying Callable doesn't take max retries from current context; uses defaults instead
> --------------------------------------------------------------------------------------
>
> Key: HBASE-21193
> URL: https://issues.apache.org/jira/browse/HBASE-21193
> Project: HBase
> Issue Type: Bug
> Reporter: stack
> Priority: Major
>
> This makes it hard to change retry count on a read of meta for instance.
> I noticed this when trying to change the defaults for a meta read. I made a customer Connection inside in the master with a new Configuration that had rpc retries and timings upped radically. My reads nonetheless were finishing at the usual retry point (31 tries after 60 seconds or so) because it looked like the Retrying Callable that does the read was taking max retries from defaults rather than reading the passed in Configuration.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)