On Wed, Jan 16, 2013 at 1:30 AM, Gianluca Cecchi wrote:
Can it be that the clustered VGs on other LUNs that are skipped is
the
cause?
BTW: tomorrow I should have a SAN guy able to mask them ...
It seems it was a problem/matter related with multipath layer, and I could
create the SD without masking the clustered LUNs
In fact, as I blacklisted the disks (PVs) that composed the VGs of the
interfering clustered logical volumes I was able to add the storage domain
Added this in my node multipath.conf
blacklist {
wwid 3600507630efe0b0c0000000000001183
wwid 3600507630efe0b0c0000000000001082
wwid 3600507630efe0b0c0000000000001182
wwid 3600507630efe0b0c0000000000001081
wwid 3600507630efe0b0c0000000000001181
wwid 3600507630efe0b0c0000000000001080
wwid 3600507630efe0b0c0000000000001180
}
In case you think it should have succeeded also without multipath
blacklisting, you can compare previous vdsm.log with errors I already sent
with the one generated right now with successful creation here:
https://docs.google.com/file/d/0BwoPbcrMv8mvS0NKd2FSWl9Yb0k/edit
Gianluca