Cluster lost quorum

[TechLineX]

Should I increase it at all 3 hosts? When hosting about 200 LXC-Containers & 60KVM-Containers on these 3 Hosts, how to increase these?


fs.inotify.max_queued_events
fs.inotify.max_user_instances
fs.inotify.max_user_watches

Regards

 

[TechLineX]

Think they got the same problem:
https://lists.linuxcontainers.org/pipermail/lxc-users/2014-August/007611.html

Same here:
https://forum.proxmox.com/threads/pve-kills-all-containers-due-to-out-of-memory.29115/

I also store all 150LXC-Servers at 2x 2TB WD Gold (ZFS-RAID1) and share it with a NFS-Share to the other hosts.

Could it also hang together with this?
https://forum.proxmox.com/threads/zfs-swap-crashes-system.25208/

 

[TechLineX]

sounds like you should bump your inotify limits: https://bugzilla.proxmox.com/show_bug.cgi?id=1042

Did that, After 9 days I'm getting the same Problem.
First the Login to the Server isn't possible and After a Short Time the Server doesn't answers anymore. Server is hangig completely. Only Solution is a reboot.

 

[TechLineX]

Proxmox-Team?

 

[Alwin]

What does your performance monitoring report? I guess there musst be some load on the machine, that could be a reason for services timing out.

In the section of the docs you find some options to limit RAM and/or swappines for ZFS.
https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_zfs_on_linux