Stability issues with LXC containers running

DzAirmaX

Member
Sep 8, 2016
10
3
23
Hi team,

First, thanks for your work. Proxmox looks like a robust project.

I am encoutering some stability issue with my server: I thought first it was link to my Hardware (like a bad HDD or a faulty power supply).

But all of the tests came back good (Memtest OK, SMART OK, stress test by recompiling kernel OK). What I noticed is that the host will be rock solid only if no LXC are launched ...

After launching lxc containers, after approximately 2 days, the system hangs. No log are available, I did install Kerndump and mcelog but impossible to get information on the reason why it's crashing. The screen is frozen on the login and the host is unreachable.

I can provide more details if you need those, please help I am getting out of ideas ...
 
A few questions:
  • How many containers are you launching?
  • Are they resource intensive? i.e. eating up memory/CPU
  • Is it a clustered setup?
  • Free version or supported version?
  • Is this a fresh install?
  • Have you tried running VMs as opposed to CTs?
Please provide hardware specs and container configuration values (screenshots with any IP or other identifiable info blurred out would help).
 
Ok, I finally found the solution!

The problem was linked to the CPU bios settings: CPU C-State were a really bad idea... Now everything is stable!

PS: I deleted my previous posts for clarity