activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Hiram Chirino <hi...@hiramchirino.com>
Subject Re: Activemq 5.9 leveldb replication issue
Date Mon, 03 Jun 2013 15:54:16 GMT
Hi Freddy,

That 'short record at position' is odd.  Basically it means it tired
to read a message from the log file and the read call read fewer bytes
than what we requested.  I guess that's valid for an OS to do and we
just need to issue an additional read to load the remaining bytes.
I've just committed a change that should help with that:
http://is.gd/0nMmLH

So try out a new snapshot and keep an eye for that exception and let
me know you see it happen on your system again.

Thanks!


On Fri, May 31, 2013 at 1:28 PM, heimdull <freddy@cfandersen.com> wrote:
> I tried this morning to replicate the issue on these two servers and was not able to
do so. The only issue that I got was a few messages that got stuck in the queue with a consumer
holding them hostage. either restarting that consumer or moving them to a different queue
and back resolved that.
>
> I did look at older errors that I had and found there recovery mode errors:
>
> INFO   | jvm 1    | 2013/05/19 12:15:22 |  INFO | DB recovered from failure.
> INFO   | jvm 1    | 2013/05/19 12:15:22 |  WARN | DB operation failed. (entering recovery
mode)
> INFO   | jvm 1    | 2013/05/19 12:15:22 | java.io.IOException: short record at position:
2467052 in file: /opt/activemq/data/0000000000000000.log, offset: 2467052
> INFO   | jvm 1    | 2013/05/19 12:15:22 |       at org.apache.activemq.leveldb.RecordLog$LogReader.read(RecordLog.scala:287)


--
Hiram Chirino

Engineering | Red Hat, Inc.

hchirino@redhat.com | fusesource.com | redhat.com

skype: hiramchirino | twitter: @hiramchirino

blog: Hiram Chirino's Bit Mojo

Mime
View raw message