Tim Cook
2006-Dec-22 22:50 UTC
[zfs-discuss] Re: Difference between ZFS and UFS with one LUN froma SAN
More specifically, if you have the controllers in your array setup in an
active/passive setup, and they have a failover timeout of 30 seconds,
and the hba''s have a failover timeout of 20 seconds, when it goes to
failover and cannot write to the disks... I''m sure *bad things* will
happen. Again, I haven''t tested this scenario, but I can only imagine
it''s not something that can be/should be/is recovered from gracefully.
--Tim
-----Original Message-----
From: Robert Milkowski [mailto:rmilkowski at task.gda.pl]
Sent: Friday, December 22, 2006 3:18 PM
To: Jason J. W. Williams
Cc: Tim Cook; zfs-discuss at opensolaris.org; Shawn Joy
Subject: Re[2]: [zfs-discuss] Re: Difference between ZFS and UFS with
one LUN froma SAN
Hello Jason,
Friday, December 22, 2006, 5:55:38 PM, you wrote:
JJWW> Just for what its worth, when we rebooted a controller in our
array
JJWW> (we pre-moved all the LUNs to the other controller), despite using
JJWW> MPXIO ZFS kernel panicked. Verified that all the LUNs were on the
JJWW> correct controller when this occurred. Its not clear why ZFS
thought
JJWW> it lost a LUN but it did. We have done cable pulling using
ZFS/MPXIO
JJWW> before and that works very well. It may well be array-related in
our
JJWW> case, but I hate anyone to have a false sense of security.
Did you first check (with format for example) if LUNs were really
accessible? If MPxIO worked ok and at least one path is ok then ZFS
won''t panic.
--
Best regards,
Robert mailto:rmilkowski at task.gda.pl
http://milek.blogspot.com