The Proxmox host will reboot, how to find the reason?

Question

linux

The Proxmox host will reboot about once a day, at night.

last reboot
reboot system boot 4.4.19-1-pve Wed Jul 12 06:09 - 14:56 (08:47)
reboot system boot 4.4.19-1-pve Tue Jul 11 ​​02:15 - 14:56 (1+ 12:40)
reboot system boot 4.4.19-1-pve Mon Jul 10 05:39 - 14:56 (2+09:16)
reboot system boot 4.4.19-1-pve Sun Jul 9 06:04 - 14: 56 (3+08:51)
reboot system boot 4.4.19-1-pve Sat Jul 8 05:12 - 14:56 (4+09:43)
reboot system boot 4.4.19-1-pve Fri Jul 7 06: 13 - 14:56 (5+08:42)
reboot system boot 4.4.19-1-pve Thu Jul 6 04:53 - 14:56 (6+10:02)

I didn't find anything in the logs myself. Sinned on iron. Changed everything except the disks. Drives judging by smart are in order.
Linux 4.4.19-1-pve #1 SMP x86_64 GNU/Linux
ZFS raid1

Answer 1

Most likely your problem is that ZFS and its L2ARC cache eat up all the memory.
Then you need to set the cache value to the minimum possible, for example, 2gb.
https://www.solaris-cookbook.eu/linux/linux-ubuntu...
Another point is that if your HP server can work wachdog at some point.
You can disable GRUB_CMDLINE_LINUX_DEFAULT="quiet nmi_watchdog=0
If it helps, you need to understand further why it works.
https://forum.proxmox.com/threads/random-restartin...

Answer 2

Does the problem reoccur when Cron is disabled? Is there anything else in the logs?
I would suggest you send dmesg over the net to see. what was "in life" a second before the fall.

Answer 3

I had a similar problem in proxmox with the openvs kernel on the sandbridge, it only helped to move to another server, not the sandbridge,
alas, no console or some other things showed nothing, very similar to the same case

Answer 4

Hello
If the CPU is Intel 6 or 7 generation then look here
https://lists.debian.org/debian-devel/2017/06/msg0...