Memory leak while using LXD on Ubuntu 18.04

I’ve got 2 VM’s - one running Ubuntu 16.04 and another running Ubuntu 18.04. Both VM’s are configured identically in that they both have 8 cores and 32GB RAM and are both running similar workloads.

On the 16.04 VM (where I have 36 containers running - a mix betrween 16.04, 18.04, and CentOS 7), RAM usage never exceeds ~16GB:

On the 18.04 VM (where I have 26 containers running - all 18.04 containers), RAM usage continuously climbs to the point where SWAPing eventually kills the server and I have to reset to recover:

I’ve got memory limits turned on for all containers on both VM’s but it appears 18.04 doesn’t honor that.

Please let me know what else I need to add to this post to help in getting to the root cause.

I think it becomes from the programs that you run in the container and not from the container itself

Maybe you did not read through my entire post. Like I said, the processes running in the containers are similar on both platforms so why is it that on Ubuntu 18.04 I see RAM usage that keeps on growing until it uses all RAM?

Sorry, I can’t help you and I think no one can help you
because you didn’t specify the programs used in your Containers

If I said it was a mix of proprietary applications, some written in golang and others in nodejs, that make use of mariadb, mongodb, redis-server, and java (specifically openjdk-8), would that help? Remember, it is the same versions of these proprietary applications that are running on both platforms.

The main difference I see here is that the base operating system has changed and hence things such as memory management. I would have thought that stating what was running in the containers wasn’t terribly important but obviously not.

So, with this new information, is it now possible to get some help?