Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 24 Aug 2009 15:57:37 +0300
From:      Panagiotis Christias <p.christias@noc.ntua.gr>
To:        "Mikhail (Plus Plus)" <m@plus-plus.su>
Cc:        freebsd-fs@freebsd.org
Subject:   Re: need help with ZFS
Message-ID:  <20090824125737.GA92643@noc.ntua.gr>
In-Reply-To: <4A927CB3.3040402@plus-plus.su>
References:  <4A927CB3.3040402@plus-plus.su>

next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Aug 24, 2009 at 03:42:43PM +0400, Mikhail (Plus Plus) wrote:
> Greetings gentlemen,
> 
> I need your help with ZFS -- how I can diagnose, debug ZFS crashes, and 
> if possible make it more stable. We're running raidz pool and we're 
> having hard time to get it running smooth -- ZFS simply crashes as soon 
> as we put some load on it.
> 
> Couple months back we've built custom server to replace our old storage 
> system. New server has 7 x 1GB SATA drives, Intel Q6600 Quad core CPU, 
> 8GB RAM. OS: FreeBSD 7.2-RELEASE-p2 amd64, stock GENERIC kernel.
> 
> So I decided to try ZFS, and I was amazed by it's features. I've read 
> ZFS wiki page and my loader.conf follows:
> 
> zen# cat loader.conf
> vm.kmem_size="1536M"
> vm.kmem_size_max="3072M"
> vm.pmap.shpgperproc="1024"
> vfs.zfs.arc_min="256M"
> vfs.zfs.arc_max="384M"
> vfs.zfs.vdev.cache.size="50M"
> vfs.zfs.prefetch_disable="1"
> kern.maxproc="20000"
> zen#
> 
> and
> 
> zen# zpool status
>    pool: datapool
>   state: ONLINE
>   scrub: none requested
> config:
> 
>          NAME        STATE     READ WRITE CKSUM
>          datapool     ONLINE       0     0     0
>            raidz1    ONLINE       0     0     0
>              ad24    ONLINE       0     0     0
>              ad8     ONLINE       0     0     0
>              ad18    ONLINE       0     0     0
>              ad20    ONLINE       0     0     0
>              ad22    ONLINE       0     0     0
>              ad10    ONLINE       0     0     0
>          spares
>            ad26      AVAIL
> 
> errors: No known data errors
> zen# zpool list
> NAME                    SIZE    USED   AVAIL    CAP  HEALTH     ALTROOT
> datapool                5.44T   3.54T   1.90T    65%  ONLINE     -
> zen#
> 
> 
> Problem starts as soon as we put some load on FS - e.g. last night I 
> tried to rsync 1TB of different files from above ZFS pool to single hard 
> drive  (partitioned UFS2), but server crashed and restarted after 
> copying ~700GB of data. I also tried to stress-test it by running 100 
> torrent downloads (using rtorrent), and server also crashes and restarts 
> after running for about 30-40 minutes.
> My guess these crashes happen due to big load on filesystem - ZFS eats 
> all available memory and then server simply crashes.
> 
> Right now I'm completely lost - I can't even copy 1TB from ZFS to 
> another partition..
> How can I diagnose the issue? Is there anything available to make ZFS 
> more stable?

I would suggest you try FreeBSD 8.0 that includes the latest version of
ZFS (version 13), which fixed several problems present in 7.x. Then, 
check for crash dumps (see dumpon(8)), collect any available info and
sent it to the list.

Regards,
Panagiotis


-- 
Panagiotis J. Christias    Network Management Center
P.Christias@noc.ntua.gr    National Technical Univ. of Athens, GREECE



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090824125737.GA92643>