Jira 8.4.1 Server - GC Allocation Failure with adjusted settings

[bryon.lewis@etrainetc.com]

I wanted to run a personal Jira Instance and attempted the installation of 8.4.1 on AWS using a free tier.  Running was slow and there were a lot of GC Allocation Failure errors.  I've adjusted the JVM Max to make sure it was below the available memory and the same issue.   In the setenv.sh I lowered it to 768MB the minumum and also tried ~900MB but the same issues kept persisting. 

This isn't on an upgraded instance with hundreds of users, it is a brand new install.  I attempted it initially on RHEL, then on the Amazon AMI Linux.  I then even tried a Google Cloud Compute Engine of Ubuntu with 1.7GB of RAM and was still getting similar issues.  I don't know if this is a known bug in 8.4.1 or not.  If it is I hope this gets to the proper people.  I'm going to try 8.3.2 on an AWS instance and see if that will work better.  

Sorry if this is in the wrong place for this type of issue.

UPDATE:  Just to let you know on Google Cloud with the more RAM it would get further, but would still eventually wind up with GC Allocation Errors.  I'm assuming the addition 700MB of RAM was the cause of allowing me to get further.  It completed setup and got me to the login screen but then became almost unusable afterwards, constantly hanging up and even locking the server.

[bryon.lewis@etrainetc.com · Accepted solution]

OKAY Major update.  I got 8.3.2 running on an AMI Linux Instance.  Through looking at some logs, it looks like main issue that the old limit of 768M for a server installation is correct but also relies on a swap partition.  I believe that both Google Cloud Engine and AMI's don't initialize with a default swap partition for the extra memory usage.  Although it does seem weird that 8.4.1 even with 1.75GB of RAM still had some issues.  So you can run Jira off a free tier of Amazon, it just requires a little more configuration.