Re: replicatedLevelDB errors after failover

Was this ever resolved? I'm testing a leveldb cluster and was performing some fail over and I now have the same problem. -- View this message in context: http://activemq.2283324.n4.nabble.com/replicatedLevelDB-

Re: replicatedLevelDB errors after failover

The leveldb seems to be corrupted after 10 failovers, broker was not able to load records from leveldb, similar to the issue described in early post but now i see these additional error msg -- View this message in context: http://activemq.2283324.n4.nabble.com/replicatedLevelDB-errors-after-fai

Re: replicatedLevelDB errors after failover

I suspect those 'Could not load message seq' warnings can be ignored. Can you verify that there has been no message loss? On Fri, Dec 6, 2013 at 12:19 PM, kal123 wrote: > Tried testing with dec. 5 snapshot and getting following errors after about > 10 failovers: > > 2013-12-06 12:06:39,673 | WARN

Re: replicatedLevelDB errors after failover

Tried testing with dec. 5 snapshot and getting following errors after about 10 failovers: 2013-12-06 12:06:39,673 | WARN | Could not load message seq: 45760 from DataLocator(630f8b1, 2262) | org.apache.activemq.leveldb.LevelDBClient | ActiveMQ NIO Worker 2 2013-12-06 12:06:39,673 | WARN | No rea

Re: replicatedLevelDB errors after failover

Ran failover test again and after about 25 or so fail overs I got the below errors on the master. I did not see any other errors. I will try to run with debug turned on next week to see it helps with debugging. Note: even after this error i was able to do few more failover before things stopped w

Re: replicatedLevelDB errors after failover

Yep that helps a little. So that seems like an error on the slave. Where there any errors on the master? On Mon, Nov 18, 2013 at 5:37 PM, kal123 wrote: > > Today after the exception occurred, I updated the activemq start script to > redirect output to file instead of null and at the start i see

Re: replicatedLevelDB errors after failover

Today after the exception occurred, I updated the activemq start script to redirect output to file instead of null and at the start i see the following error. Does this help? INFO | Attaching... Downloaded 295.12/295.12 kb and 4/4 files INFO | Attached java.io.IOException: invalid record positi

Re: replicatedLevelDB errors after failover

I believe there was no other errors prior to that exception, i will try to reproduce and update you on it. I downloaded the source yesterday form git: parent: e57aeb3) | patch See AMQ-4886. Updated tearDown so it can't hang, reduced timeouts, updated to JUnit4 -- View this message in con

Re: replicatedLevelDB errors after failover

Hi that exception should have also printed out a cause. Could you include that too? If you built it yourself, which git commit did you build? If you downloaded a snapshot whats the file name of the download? On Fri, Nov 15, 2013 at 8:46 AM, kal123 wrote: > I am using the latest changes from tr

Re: replicatedLevelDB errors after failover

the issue seems to happen when broker that is being stopped generated the following exception. Store getting corrupted due to the way the stop is handled? 2013-11-15 09:26:52,881 | WARN | Transport Connection to: tcp://10.44.173.146:47163 failed: java.io.IOException: Connection reset by peer | o