accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-3321) Log exceptions in ThriftScanner#getBatchFromServer as warnings/errors
Date Mon, 10 Nov 2014 15:44:33 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-3321?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14204904#comment-14204904
] 

ASF GitHub Bot commented on ACCUMULO-3321:
------------------------------------------

Github user arielvalentin commented on the pull request:

    https://github.com/apache/accumulo/pull/18#issuecomment-62402671
  
    @joshelser I agree that because the client may recover from the is strictly not an error
case, however I feel it is good to have breadcrumbs available when trying to track down problems.
In my specific case we run into some really slow scans/writes and I don't know what the root
cause is because I am not logging at debug level. I look at the monitor UI and Accumulo is
humming along but then have to bump the client side log level in order to see that the first
attempt fails on every scan but the second attempt works just fine. 
    I find that the client side logging of first attempt failures at warn level, as opposed
to debug, to be very helpful because it allows our monitoring tools to alert our administrators
that our application may be having connectivity issues. That allows them the ability to write
reports and see how often these failures occur and potentially track down the root cause.




> Log exceptions in ThriftScanner#getBatchFromServer as warnings/errors 
> ----------------------------------------------------------------------
>
>                 Key: ACCUMULO-3321
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-3321
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: client
>            Reporter: Ariel Valentin
>            Priority: Minor
>
> While trying to debug an issue with a client that seemed unresponsive I increased the
log lever to debug and found that we were experiencing DNS failures:
> {code}
> 22:14:53,549 DEBUG [org.apache.zookeeper.ClientCnxn] (http-/192.168.220.198:8080-1-SendThread(ZKHOST:2181))
Reading reply sessionid:0x1498bc00431010f, packet:: clientPath:null serverPath:null finished:false
header:: 414,3  replyHeader:: 414,107374188046,0  request:: '/accumulo/4663a372-d7f0-4956-a110-ff063c781ead/tservers/TSERVER_HOST:10011/zlock-0000000036,T
 response:: s{107374187891,107374187891,1415398174207,1415398174207,0,0,0,92758924158107923,33,0,107374187891}

> 22:14:53,922 DEBUG [org.apache.zookeeper.ClientCnxn] (http-/192.168.220.198:8080-1-SendThread(ZKHOST:2181))
Reading reply sessionid:0x1498bc00431010f, packet:: clientPath:null serverPath:null finished:false
header:: 415,4  replyHeader:: 415,107374188046,0  request:: '/accumulo/4663a372-d7f0-4956-a110-ff063c781ead/tservers/TSERVER_HOST:10011/zlock-0000000036,T
 response:: #54534552565f434c49454e543d7330322d6265732d646e6f6436323a3130303131,s{107374187891,107374187891,1415398174207,1415398174207,0,0,0,92758924158107923,33,0,107374187891}

> 22:14:53,936 DEBUG [org.apache.accumulo.core.client.impl.ThriftScanner] (http-/192.168.220.198:8080-1)
Error getting transport to TSERVER_HOST:10011 : org.apache.thrift.transport.TTransportException:
java.net.UnknownHostException
> {code}
> There is another exception that is logged at debug in getBatchFromServer and I would
like these messages to log at WARN or ERROR level. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message