zfs discuss - Nov 2007 - Solaris 8/07 Zfs Raidz NFS dies during iozone test on client host

Hi,

Well I have a freshly built system with ZFS raidz.

Intel P4 2.4 Ghz
1GB Ram
Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller
(2) Intel Dual Port 1Gbit nics

I have (5) 300GB disks in a Raidz1 with Zfs.

I''ve created a couple of FS on this.

/export/downloads
/export/music
/export/musicraw

I''ve shared these out as well.

First with ZFS ''zfs set sharenfs=on''

Then with /etc/dfs/dfstab.

On a Solars 10 client host I would mount that Files System Vers=3.
Then CD to that mountpoint and proceed to run iozone {from Blastwave}

<root at vmserver2 /mnt/downloads>
# iozone -Racb iozone.`hostname`.wks > iozone.data

Sometime during the test I get a ''NFS server timsnas-nfs not responding
still trying''
NFS is dead at this point. To no avail. A reboot of the NFS server is order to
recover.

The clients mount this share:
timsnas-nfs:/export/downloads -       /mnt/downloads  nfs   - yes    
vers=3,bg,rsize=8192,wsize=8192,timeo=14,proto=tcp,intr

The Server is set to NFSv3 only:
# Sets the minimum version of the NFS protocol that will be registered
# and offered by the server.  The default is 2.
NFS_SERVER_VERSMIN=3

# Sets the maximum version of the NFS protocol that will be registered
# and offered by the server.  The default is 4.
NFS_SERVER_VERSMAX=3

At this point I have no clue what the issue is....

On the NFS server I''ve successfully ran ''iozone'' 2X.
I''ve also ran scrub on this zpool.

Please advise...

Tim
 
 
This message posted from opensolaris.org