zvol + raidz issue?

[Niki Hammler]

Am 26.08.2012 22:13, schrieb Freddie Cash:
> (Sorry for top-post, sending from phone.)
> 
> Please show the command-line used to create the zvol. Especially the
> recordsize option.  When using zvols, you have to make sure to match the
> recordsize of the zvol to that of the filesystem used above it.
> Otherwise, performance will be atrocious.

Hi,

Sorry for my third posting on this.
Now I strictly followed your suggestion and used

zfs create -b 128k -V 500g plvl5i0/zvtest

(with 128k being the recordsize of the dataset in the zpool).

Suddenly the write performance increased from the 2.5 MB/s to 250 MB/s
(or 78MB/s when using bs=4096 with dd)

1.) How can this explained?
2.) Is there any problem when choosing -b 128k (can I always blindly
choose -b 128k)?

Remember again that the problem ONLY occurs with raidz1+zvol+force 4096
block alignment and in no other case!

Regards
Niki

> On Aug 26, 2012 11:50 AM, "Niki Hammler" <mailinglists at nobaq.net
> <mailto:mailinglists at nobaq.net>> wrote:
> 
>     Hi,
> 
>     Given: new HP Proliant Microserver N40L (4 GB RAM) and 3x2TB SATA drives
>     (SAMSUNG HD204UI, ST32000542AS, WDC WD20EARX-00PASB0).
> 
>     Goal: RAIDz1 containg datasets and zvols to be exported via iSCSI.
> 
>     Issue: When I create a zvol on a RAIDz1 I get horrible performance (few
>     MB/s or less).
> 
>     First test: 500G zvol on a mirror (freshly created):
> 
>     # zpool list
>     NAME      SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
>     plvl1i0  1.81T  1.97G  1.81T     0%  ONLINE  /mnt
>     # zfs list
>     NAME               USED  AVAIL  REFER  MOUNTPOINT
>     plvl1i0            500G  1.30T   112K  /mnt/plvl1i0
>     plvl1i0/zvtest     500G  1.78T  1.97G  -
>     # dd if=/dev/zero of=/dev/zvol/plvl1i0/zvtest bs=2048k count=1000
>     1000+0 records in
>     1000+0 records out
>     2097152000 bytes transferred in 17.318348 secs (121094230 bytes/sec)
>     #
> 
>     Corresponds to 115,48 MB/s which is good (similar results for a single
>     drive).
> 
>     Second test: 500G zvol on the 3x2TB raidz1 (freshly created):
> 
>     # dd if=/dev/zero of=/dev/zvol/plvl5i0/zvtest bs=2048k count=1000
> 
>     1000+0 records in
>     1000+0 records out
>     2097152000 bytes transferred in 700.126725 secs (2995389 bytes/sec)
>     #
> 
>     which is only 2,85 MB/s.
> 
>     Remark: Both pools are created with the force 4096 alignment option
>     (since I have 512 and 4096 drives mixed).
> 
>     Now is the point where you might say the problem is related to the
>     raidz1. But it is not: I created a 500G dataset in the same RAIDz pool
>     and copied about 100G data onto it with rsync+ssh. Result: about 28MB/s
>     end2end performance which is reasonable.
> 
>     Are there any issues with zvol + raidz1? Google resulted in empty result
>     set.
> 
>     I run a minimal FreeBSD 8.2 (FreeNAS):
> 
>     # uname -a
>     FreeBSD zetta 8.2-RELEASE-p9 FreeBSD 8.2-RELEASE-p9 #0: Thu Jul 19
>     12:39:10 PDT 2012
>     root at build.ixsystems.com:/build/home/jpaetzel/8.2.0/os-base/amd64/build/home/jpaetzel/8.2.0/FreeBSD/src/sys/FREENAS.amd64
>      amd64
> 
>     Regards,
>     Niki
> 
> 
>     PS: This is also posted on
>     http://forums.freenas.org/showthread.php?p=35590
>     _______________________________________________
>     freebsd-fs at freebsd.org <mailto:freebsd-fs at freebsd.org> mailing list
>     http://lists.freebsd.org/mailman/listinfo/freebsd-fs
>     To unsubscribe, send any mail to "freebsd-fs-unsubscribe at freebsd.org
>     <mailto:freebsd-fs-unsubscribe at freebsd.org>"
>