Advice on kernel panics

[Raimo Niskanen]

On Thu, Jun 01, 2017 at 11:58:48AM +0200, Andrea Venturoli wrote:
> On 05/29/17 11:20, Raimo Niskanen wrote:
> > Hello list.
> > 
> > I have a server that panics about every 3 days and need some advice on how
> > to handle that.
> 
> As Doug suggested, it's probably an hardware fault.
> 
> 
> 
> However, I've many times seen boxes crash every day or two when an 
> UFS+SU filesystem gets corrupted. Background fsck won't detect this.
> 
> In order to stop the crashes, I always had to boot into single user mode 
> and do a fsck -y (twice if asked) on every UFS+SU filesystem.
> 
> Furthermore, if you are using SU+J, you'll have to run fsck first to 
> deal with the journal and then fsck it again (ignoring the journal).
> 
> This has brought several boxes I manage back on track on many occasions.
> 
> You might want to try this, since the only downside is downtime.

Good advice, but I should have given more details from the start.
I have only ZFS and NFS filesystems (and devfs andalso fdescfs).

-- 

/ Raimo Niskanen, Erlang/OTP, Ericsson AB