Hi guys,

I'm running:
Version: 1.3.2 (e7c49b6) (Hansa)
JVM: PID 2039, Oracle Corporation 1.7.0_80 on Linux 3.2.0-4-amd64

On an EC2 c4.2xlarge box with >=1000 IOPS EBS allocated to GL. I had it 
running on a c4.4xlarge as well, with all cores and memory-sensitive params 
updated, but it didn't make any difference.

Important to mention: it's an all-in-one setup (MongoDB, ES, GL on one box).

Nothing relevant to the problem is shown in the logs.

The behaviour is something of a cliche as far as I could tell by reading GH 
issues and mailing list threads:
1. 'Process buffer' gets filled.
2. Messages keep coming in, but they're not going out (read processed) 
anymore.
3. Journal fills up and never recovers.

JVM heap memory is set to 2.9GB and seems to be behaving well, even when GL 
stops processing messages.

The only fix that I have at the moment is to restart the graylog-server 
service. Then the journaled messages begin to be processed, the journal is 
being emptied and eventually the 'Process buffer' is emptied as well.

I tried removing the contents of the *journal/* dir, but to no avail.

On another note, I have to manually trigger the recalculation of the 
indices ranges once the service is up again.

I'd appreciate any hints.

Thanks.

-- 
You received this message because you are subscribed to the Google Groups 
"Graylog Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/graylog2/d15b5442-76fa-4eec-9315-e127fdec8a1e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to