You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hbase.apache.org by "Adela Maznikar (JIRA)" <ji...@apache.org> on 2014/05/27 05:05:02 UTC

[jira] [Created] (HBASE-11254) [0.89-fb] Improve retry logic in HTableMultiplexer.HTableFlushWorker

Adela Maznikar created HBASE-11254:
--------------------------------------

             Summary: [0.89-fb] Improve retry logic in HTableMultiplexer.HTableFlushWorker
                 Key: HBASE-11254
                 URL: https://issues.apache.org/jira/browse/HBASE-11254
             Project: HBase
          Issue Type: Improvement
    Affects Versions: 0.89-fb
            Reporter: Adela Maznikar
            Assignee: Adela Maznikar
             Fix For: 0.89-fb


The failed puts in HTableMultiplexer.HTableFlushWorker are immediately resubmitted in the original queue. This queue is being flushed every 100ms (default configuration) and the maximal retry of a failed put is by default 10 times (client configurable). That leaves us with ~1 second to complete a Put (with default configuration). We can improve this by gradually increasing the time for retrying, each time when the put fails. (example: first time retry after 100ms, then after 200ms, 400ms, etc). 



--
This message was sent by Atlassian JIRA
(v6.2#6252)