iSCSI has built-in 'ping' implemented with NOP-{IN/OUT} protocol
messages. This heartbeat is initiator-driven, so you should check
whether this is configurable with iSCSI initiator you use and choose
appropriate setting.Second part of the equation, namely how frequently zfs checks for pool disks health (in the absence of in-flight I/O to signal disk down with EIO), belongs to zfs-discuss. -- Regards, Andrey On Feb 15, 2008 11:02 PM, Engle, Victor <[EMAIL PROTECTED]> wrote: > > That is consistent with my testing as well. Also, without the latest > iscsi patch mpxio will never bring that path back online once it is > failed. > > > > > -----Original Message----- > > From: Ross [mailto:[EMAIL PROTECTED] > > Sent: Friday, February 15, 2008 2:16 PM > > To: [email protected] > > Subject: Re: [storage-discuss] iSCSI disconnect = 5 minute system hang > > > > And a bit more info. I just discovered the "time" command, > > and found out that either my reflexes are shot, or I've > > forgotten how to count... > > > > Using "time" to see how long it takes for "zpool status" to > > recover after downing an iSCSI device reports it as taking 3 > > minutes for ZFS to realize that the target is unavailable. > > > > > > This message posted from opensolaris.org > > _______________________________________________ > > storage-discuss mailing list > > [email protected] > > http://mail.opensolaris.org/mailman/listinfo/storage-discuss > > > _______________________________________________ > storage-discuss mailing list > [email protected] > http://mail.opensolaris.org/mailman/listinfo/storage-discuss > _______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
