googol

I managed to find my way to the unraid webgui repository on github which seems to host the init script that does the mounting. I've opened an issue there to see if I could find an answer there: https://github.com/unraid/webgui/issues/1598

Thanks for the reply @dlandon . Do you have an idea of who might know more about this topic?

Hi @dlandon, are you familiar with the cgroup mounts on unraid?

@JorgeB sorry to pull you in like this out of the blue, but I see you are an active moderator. It seems to me that this forum might not be the best place to ask such a specific question about the OS. Do you know who I should contact about this, or who could be pulled in to this thread to shine a bit of light on this?

I've been running Hashicorp Nomad to control the docker containers on my Unraid server for a while now, and encountered some errors when updating to the 1.7.x versions of it. The 1.7.0 version had some major updates to the way they handle cgroup version support detection, and is now failing to identify the cgroup v2 support of Unraid. This seems to be caused by a rather odd mount configuration on Unraid. For the path `/sys/fs/cgroup` there are two mounts on top of each other: # mount -l | grep cgroup cgroup_root on /sys/fs/cgroup type tmpfs (rw,relatime,size=8192k,mode=755,inode64) none on /sys/fs/cgroup type cgroup2 (rw,relatime) These file systems are mounted in /etc/rc.d/rc.S. Comparing it to the init script in the upstream slackware sources, it seems to be different: https://mirrors.slackware.com/slackware/slackware64-15.0/source/a/sysvinit-scripts/scripts/rc.S. Can someone confirm this is a unraid-specific change? Is there some specific reason for the tmpfs mount, which seems to be unnecessary and causing my problem. I've created a ticket on the Nomad issue tracker, link here for reference: https://github.com/hashicorp/nomad/issues/19735