Server crashing with kernel panic and kernel taint line
I'm running a Proxmox server (underlying system is Debian). The server tends to crash at random time intervals- sometimes it goes a week or so without crashing, and sometimes it takes several weeks. The system is completely unresponsive at this point.
On the screen it looks like the memory is dumped and I noticed the mention of a service being "tainted." See this image After a little research I learned to run the following command to determine the kernel taint value: Code:
root@proxmox:~$ cat /proc/sys/kernel/tainted |
On the crash: The main lines of interest has rolled off screen. Often Shift_PgUp gets it back.
On the 'taint': Linux kernel maintainers insist on code of what they consider a 'high' standard from experienced coders. When outside modules are inserted (e.g. AMD or NVIDIA graphics drivers) OSS purists like Debian will consider the kernel tainted. A grep of syslog will probably enlighten you, and a search for independently compiled modules in /lib/modules/<version>/misc will certainly narrow the field. They're probably right; External module code is often written by guys who can't get code into the main kernel, and they are sometimes more buggy and less thoroughly tested. The kernel supports about 2 dozen architectures, so they need to be strict. |
A simple seach should have found this page from the kernel doco. An indication of "Tainted" doesn't cause crashes - but the kernel devs won't deal with a crash involving a proprietary module as in this case.
zabbix is mentioned, but it might just be a victim. Or it might be the cause. As stated, not enough to go on, but I'd start by looking for high memory consumers, and anything spinning a CPU. |
All times are GMT -5. The time now is 09:01 AM. |