Hi Paolo,
Could you say more about the kind of error you are seeing?
You mention I/O error "without any message in console or in messages",
so where are you seeing these errors, on the ZFS filesystem? Or are
those seen at the output of the clquorum command? What exactly is the
error message and at the point when you execute the command, what is
the status of the Zpool?
I do know that there is an SC issue with adding a quorum disk
into a Zpool, because ZFS reformats the disk with EFI label and that
looses the quorum information on the disk. However in your case you
are adding the disk as a quorum device AFTER it has already been
added to the Zpool, that is supposed to work.
Is d4 > 1 TB by any chance? IIRC, there were some issues there.
The more info about the specific error scenario and error message you
share, more helpful it would be.
Regards,
-ashu
Paolo Merisio wrote:
> Ok,
>
> I check for fencing used on shared disk (same Backends exported multiple time)
>
> global_fencing: nofencing
>
> default_fencing: nofencing for all did device.
>
>
> I use d4 = /dev/rdsk/c0d1 in zpool pool1
> Then I create HAStorage resource to manage this pool and all work fine.
> (switch to all 3 node without problem)
>
> If I try to use d4 as quorum device I receive I/O error without any message
> in console or in messages.
>
> Any idea?
> Thanks
