Thanks. I installed pp_hds_modular v2.00 and updated multipath.conf to use
the queue_if_no_path option.
I'm not using Qlogic hba's, but Emulex LPe11000-S instead. Could this be
part of the problem?
Btw, multipath -ll reported the failed path as back online before the
second controller on the DF600 was taken offline. We also have a Solaris 9
machine running storage foundation with dmp connected to the df600 and it
didn't have any issues. The sf dmp took longer to detect the path was back
online, so we waited longer for it to enable the path before taking the
second controller offline.
On Tue, 24 Jul 2007, Matthias Rudolph wrote:
Sometimes it happens that both pathes to a SAN LUN are not available for
a short time. Because it is not possible to enable any timeouts inside
the HBA driver (QLogic) you have to handle this situation a layer above.
You can do it with the DM Multipath feature "queue_if_no_path".
I have upgraded the HDS prioritizer to version 2.00. You will find all
the stuff on:
http://homepage.mac.com/i_am_fable/FileSharing36.html
Regards
Matthias Rudolph
Country Support Germany
HITACHI DATA SYSTEMS
matthias.rudolph@xxxxxxx
--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel
--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel
