Ignition 8 Garbage Collection

#1

I have been having issues with memory consumption increasing over the course of a week, eventually bringing the server to its knees. Looking at the task manager reveals that Zulu Platform x64 Architecture is using more than its allocated 2Gb of RAM; a simple reboot fixes the issue for a week or so. My server has 4Gb of RAM total, with 2GB allocated to Ignition (see .conf file excerpt below).

A few forum posts regarding 7.x say to use the G1C1 garbage collector instead of the default ‘UseConcMarkSweepGC’. Now I’m on 8.6, and the additional parameters section in my ignition.conf file doesn’t show any garbage collector that I have read about, but to be honest I don’t really know what I’m looking at:

# Java Additional Parameters
wrapper.java.additional.1=-Ddata.dir=data
#wrapper.java.additional.2=-Xdebug
#wrapper.java.additional.3=-Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=*:8000


# Initial Java Heap Size (in MB)
wrapper.java.initmemory=1024

# Maximum Java Heap Size (in MB)
wrapper.java.maxmemory=2048

I’m thinking I should just add a line with this text? I’m tempted to throw more memory at my server, but it seems to run great for a while. This must be a garbage collection issue?

wrapper.java.additional.1=-XX:+UseG1GC**
0 Likes

#2

No, it’s already using G1GC by default, you don’t need to add anything.

This is probably a memory leak of some sort, which I guess is technically a “garbage collection issue” if you squint at it, but changing your GC isn’t going to fix a memory leak and any recommendations to the contrary aren’t founded in reality.

Most of the old advice to switch from CMS to G1 was for alleviating issues with clock drift.

I’d call support and let them start looking at your project to see if there’s anything that might be causing a leak.

1 Like

#3

Here is an image that our IT guy gave me showing the memory consumption over a week
memory%20trend%2002142020

I just scheduled the server to reboot every Sunday evening when there isn’t much production, but this is just a band aid kind of.

0 Likes

#4

That looks like the graph you’d see from a classic memory leak… probably need to call support and have them get involved troubleshooting. It’s often pretty hard to track these down.

0 Likes