Confluence crash/instability

Saad Taryous September 22, 2019

Hello,

Since upgrading to confluence 6.15.7, i started to have crashs several time (more then 5 crashs) per  day.

After checking out the logs i found a timeout error caused by one of the workbox plugins, so i disabled all the workbox plugins and in-app notifications, with that the number of crash got to one crash per day.When i run the health check i have no error, and the JVM have always a 30% free memory, in the logs i can't found any errors ...

I also disabled the synchrony collaborative edit, after the upgrade since it require some change in the reverse proxy

The instance have 5K users, 260 spaces and runs with 1024Mo JVM min and 2048 JVM max (the bdd runs in a different box)

(I had set the 200 maxthreads property in the reverse proxy connector)

If anyone have any advice/tips/ideas for me or experienced the same behaviour ?   

Anything will be helpful since i can't found out the cause of this crashes ... :(

Thanks :)

 

1 answer

0 votes
DPKJ
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
September 22, 2019

@Saad Taryousat first glance memory allocated to JVM is way low for the size of your instance, please increase JVM Max to 4GB and restart and observe if instance still crashes.

Saad Taryous September 23, 2019

Even if the JVM is always 30~40% free?

DPKJ
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
September 23, 2019

@Saad TaryousMemory usage increases during site load, and indexing like activities so I would say yes.

Saad Taryous September 23, 2019

Ok, i will try to increase the JVM.

thx for your replay i will keep you updated.

DPKJ
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
September 23, 2019

Sure @Saad Taryous 

Also next time when crash happen make copy of all the logs before starting service again, this will us diagnose further.

Suggest an answer

Log in or Sign up to answer
TAGS
AUG Leaders

Atlassian Community Events