[storage-discuss] Re: Re[2]: Re: Testing the Solaris iScsi Target with the

[Nigel Smith]

Ok, I have created another iScsi target, but this time using a ZFS zvol as the backing store.
I did not use a separate real hard-drive for ZFS, but used a file on the boot (UFS) hard drive (30G IDE).

# cd /export/home
# mkfile 2g ./zfs-file
# zpool create tank /export/home/zfs-file
# df -h /tank
Filesystem             size   used  avail capacity  Mounted on
tank                   2.0G    24K   2.0G     1%    /tank
# zfs create -V 1.5g tank/vol1
# zpool list
NAME                    SIZE    USED   AVAIL    CAP  HEALTH     ALTROOT
tank                   1.98G    121K   1.98G     0%  ONLINE     -
# zfs list
NAME        USED  AVAIL  REFER  MOUNTPOINT
tank       1.50G   464M  24.5K  /tank
tank/vol1  22.5K  1.95G  22.5K  -
# iscsitadm create target -b /dev/zvol/rdsk/tank/vol1 target1
# zpool iostat
               capacity     operations    bandwidth
pool         used  avail   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
tank         364M  1.63G     21      6   610K   259K

I mounted the new target from the Microsft Windows iScsi initiator, and ran the 
SQLIOsim.exe for 10 minutes, using settings:
MaxOutstandingIO=50
TargetIODuration=200

The test run completed succesfully.  Everything worked fine - no errors.  Yes it was slow, but otherwise everything is ok for me.

What I would like to see now is the current version of the iScsi target code
integrated into the next build of snv.  If I could get a working iScsi target onto a bootable CD like BeleniX, I could do some proper stress testing on more modern PC's at my work location!
Thanks
Nigel Smith
 
 
This message posted from opensolaris.org