lucene-solr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yonik Seeley (JIRA)" <j...@apache.org>
Subject [jira] Commented: (SOLR-91) autowarming searchers can stack up
Date Fri, 22 Dec 2006 04:37:21 GMT
    [ http://issues.apache.org/jira/browse/SOLR-91?page=comments#action_12460379 ] 
            
Yonik Seeley commented on SOLR-91:
----------------------------------

A bad solution would be to kill the oldest searcher... this strategy could lead to starvation.
A better solution: allow a <commit/> to fail after a configurable threshold of the number
of warming searchers is reached.

Downsides:
 - commit currently does not fail (unless exceptional circumstances like an IOError)
 - If enabled on a master, two different clients doing a quick add+commit
   could now cause failure.  We could either just recommend it never be enabled for a master,
or
   some autocommit functionality could be added to enforce a minimum wait time between commits.


> autowarming searchers can stack up
> ----------------------------------
>
>                 Key: SOLR-91
>                 URL: http://issues.apache.org/jira/browse/SOLR-91
>             Project: Solr
>          Issue Type: Improvement
>    Affects Versions: 1.1.0
>            Reporter: Yonik Seeley
>
> If the commit frequency is high, and autowarming of caches is configured, sometimes an
overlap of warming searchers can occur.  Solr currently detects this and logs a warning message.
> While normally relatively harmless, multiple warming searchers can sometimes cause thrashing,
extending the time it takes to complete warming, and causing a chain reaction where more and
more searchers try to warm concurrently.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message