Hello
This is an early start of a thread for documenting a potential issue. I will add to it with findings accordinly.
I'm having some issues with on XOCE instance, commit 7a187. Three times this week, it has gotten quite unresponsive with network trafic. And looking at the console give clues as to why:
"TCP: out of memory -- consider tuning tcp_mem"
[image: 1788511260553-0ec99c11-994b-4a9f-abc8-66d7aa1c74b0-image.jpeg]
The weird part is that this is my most beefy XOCE instance of all.
4 cores, 8GiB, management agent 10 (Citrix), Rocky 10.2.
In troubleshooting I tried restarting the XO-server service. And this seems to have resolved the issue for now. But TBH, I was looking at several point at the same time. So it is possible that it had resolved itself (unlikely but possible) whilst I was troubleshooting and reading up on tcp_mem.
My other XOCE instances are setup identically, and are actually a little bit smaller in the amount of RAM assigned. And they show no problems.
Failing XO instance connects to three separate 1Host pools.
Working XO instance connects to 1 HA-pool with three hosts.
Installation is done by a port of https://github.com/cloudrootab/ansible_role_xoce, adjusted for Rocky10 instead of Ubuntu/Debian.
