Re: Did something change with ZFS and vnode caching?

From: Jonathan Chen <jonc_at_chen.org.nz>
Date: Mon, 11 Sep 2023 05:03:03 UTC 
Hi,

On 10/09/23 14:28, Garrett Wollman wrote:
> <<On Fri, 1 Sep 2023 01:04:56 +0200, Mateusz Guzik <mjguzik@gmail.com> said:
> 
>> zfs lock arrays are a known problem, bumping them is definitely an option.
> 
> This is the thing I tried next.  It took a few attempts (mostly I
> think due to my errors) but I'm now running with 512 (instead of 64)
> and plan to deploy 1024 soon, as the results are significant: while we
> still see significant loads and kmem pressure during the backup
> window, backups are able to complete some 5 to 8 hours sooner, and
> nfsd remains responsive.

How did you bump up the zfs lock array value? Is this a sysctl-tweakable 
value?

I'm currently running a recent version of stable/13 on a 24-CPU machine. 
It runs Postgresql 15. If I slam the system with a "pg_restore -j 20 
...", the system will freeze more often than not. This issue does not 
seem to appear if I use UFS as my backing store, so I'm left to assume 
that it's ZFS related. I'm wondering whether your tweaks will affect my 
system's behaviour.

Cheers.
-- 
Jonathan Chen <jonc@chen.org.nz>