after some frustrating experiments using open-iscsi with FC5 and Xen. I
would like to do the same tests with Core-iSCSI. I saw some success
messages with FC5 already. Is there anybody using
it together with Xen?
It also appears (to someone like me who hasn't looked very closely)
that Open-iSCSI is trying to make use of the latest kernel structures
and the ground keeps changing underneath them. They will eventually
have a really good solution, but for now it looks like a lot of careful
selections of the best rev of Open-iSCSI matched with the best rev of
the kernel plus intensive reading of mailing lists for patches is
needed for stability. (this is all an almost uneducated guess).
I would note that when you are playing in this area (that is, pretty
close to the bleeding edge) you'd better not expect "instant
gratification" from any of the solutions you pick. It's pretty easy
to go "off the page" into uncharted/untested territory.
Best wishes, good luck, and keep me posted (and I'll do the same)...
Re: [Xen-users] iSCSI-LUN as VBD
from [Peter Simon] [Permanent Link][Original]
To: Martin Heller <martin.heller@xxxxxxxxxxxxxxxxx>
Subject: Re: [Xen-users] iSCSI-LUN as VBD
From: Peter Simon <Peter.Simon@xxxxxxxxxxxxxxx>
Date: Tue, 23 May 2006 18:32:02 +0200
Martin Heller wrote:
> Peter Simon wrote:
> > at the list I read the that some guys have this setup working:
> > Access though iSCSI in dom0 to the target, starting domU directly form
> > iSCSI-LUN as vbd.
> We got xen-2.0.7 with Debian Sarge 2.6.11 and core-iscsi 184.108.40.206 as
> initator and iscsi-tools 3.3.
Thanks for this hint.
I've tested core-iscsi and it works pretty well with xen-3.0.2, SLES9,
Kernel 2.6.16 and our NetApp-Box :-)
Now I'm asking me, if their is something in core-iscsi to do persistent
LUN mapping, like this /dev/disk/by-id/...-thing in open-iscsi?
Schöne Grüße aus München ;-)
Xen-users mailing list
Server: 2.6.16-1.2289_FC6xen with iscsitarget-0.4.13
Client: FC5 (2.6.16-1.2133_FC5) and core-iscsi-v220.127.116.11
CHANNEL="0 1 eth1 192.168.1.100 3260 0"
DEVICE="0 0 1 disk0"
When I mount channel 0 and start bonnie++ the client hangs.
BUG: soft lockup detected on CPU#0!
EIP is at scsi_request_fn+0x32e/0x366 [scsi_mod]
Perhaps this is a bug in the scsi_module, which was also the problem I
had with open-iscsi.