Getting slow and jumpy ZFS write performance

openindianaopensolarisperformancesolariszfs

I have a 10-disk RAIDZ in Solaris 11.1. Its connected by 8G fibre. I have created a ZVOL, made it available with COMSTAR and mounted it on another machine connected by fibre. When I benchmark it with large files, I get ~700MB/s read and ~150MB/s write. That's averaged. Upon closer inspection the read speed is totally steady but the write speed is constantly fluctuating from all the way down from 50MB/s to 400MB/s.

I know ZVOLs use sync writes but should I get this slow write speed and be seeing this fluctuation? I feel like I should get better (or at least consistent) write without having to have a ZIL.

Best Answer

RaidZ has write performance of a single disk. It is not directly equivalent to raid5. If you want performance you need to stripe it or use mirrors.

Related Solutions

COMSTAR and ZFS zvol giving bursty write pattern

That's normal for ZFS. What's happening is it's coalescing the writes in memory and then committing them to disk at the same time. This a great wiki about what it's doing and how you can tune it:

http://www.solarisinternals.com/wiki/index.php/ZFS_Evil_Tuning_Guide

Nfs – ZFS nfsshare to export RW and RO host

I think you have a syntax problem. The value of the sharenfs property is the options that get sent to share(1). Try

zfs set sharenfs=rw=host1:host2:host3,ro=host4 pool1

unless you really do need to be exporting the filesystems with root permissions to all those hosts, in which case it's safest to be explicit:

zfs set sharenfs=rw=host1:host2:host3,root=host1:host2:host3,ro=host4 pool1

(or better yet, use a netgroup or network instead).

Best Answer

Related Solutions

COMSTAR and ZFS zvol giving bursty write pattern

Nfs – ZFS nfsshare to export RW and RO host

Related Topic