Re: curious crashes when under memory pressure

On 2025-01-04, Craig Leres <leres@freebsd.org> wrote:
> On 1/4/25 09:36, Karl Denninger wrote:
>> What's the argument for swapping to a mirror in the first place? If the 
>> issue is throughput IMHO the answer is to swap to multiple devices.
>
> I thought the idea here is if your swap is not redundant and it goes 
> away the system crashes (or otherwise misbehaves)?
>
> I used a beefy poudriere build server (48 cores, 192G, 6 x 480GB nvme) 
> and I build > 1000 packages daily. I see lots of build failures (usually 
> 2-3 a week) but instead of core dumps I see "Killed" in the poudriere 
> logs. Appended is one from today. (Not sure what's up with the symlink 
> failure at the very end...)

There might be a corresponding message in the system log for the
"Killed", that would at least tell us the signal, maybe more.

I am doing this on my temporary building bhyves:
    echo "vm.panic_on_oom=1" >> $tempdir/etc/sysctl.conf
    echo "vm.pageout_oom_seq=5120" >> $tempdir/etc/sysctl.conf
    echo "vm.oom_pf_secs=180" >> $tempdir/etc/sysctl.conf
    echo "vm.pfault_oom_attempts=-1" >> $tempdir/etc/sysctl.conf

https://gitr.daemon.contact/sysup/tree/subr_virt.sh#n148