ZFS and large directories - caveat report

[Ivan Voras]

On 21 July 2011 20:48, David P Discher <dpd at bitgravity.com> wrote:
>
> Ivan -
>
> What's your uptime ?

Currently about a day since other people rebooted the server in panic.

> Are you using l2 arc ?

As a separate device? No.

> what is the value of 'sysctl kstat.zfs.misc.arcstats.evict_skip' ?
>  is this increasing quickly ?

www:~> sysctl kstat.zfs.misc.arcstats.evict_skip
kstat.zfs.misc.arcstats.evict_skip: 4442021
www:~> sysctl kstat.zfs.misc.arcstats.evict_skip
kstat.zfs.misc.arcstats.evict_skip: 4442410

It increases (like the two values I've given) about every 5 seconds,
which would mean approx. 70 per second.

> How much cpu are the 'arc_reclaim_thread' and 'l2arc_feed_thread' taking up ?
>  top -SHb 500 | grep arc

Almost none.

   36 root       -8    -     0K    80K arc_re  0   0:41  0.00%
{arc_reclaim_thre}
   36 root       -8    -     0K    80K l2arc_  0   0:01  0.00%
{l2arc_feed_threa}