The cluster name is tg24. It consists of two nodes, tg22 and tg23.
When tg23 was the master:
tg22 would show the drive letters, but there were no files listed. (as
expected)
tg23 would show the drive letters and have access to the disks. (as
expected)
The server name tg22 would have no shares. (as expected)
The server names tg23 and tg24 would sow the cluster shares. (as expected)
When tg22 was the master: (this is the problem scenario)
tg22 would show the drive letters and have access to the disks. (as
expected)
tg23 would show the drive letters and have access to the disks. (this is
wrong) !!
The server names tg22 and tg24 would sow the cluster shares. (as expected)
The server name tg23 would have no shares. (as expected)
tg23 should not have had read/write access to the disks, but it did.
The next day, (just after I posted the first message) we began to have file
corruption on the shared disk.
Currently tg22 is off and tg23 is the master. This will be in place until I
get a clue as to the problem or have time off-hours to do more tests.
The shared disks are an iSCSI RAID array. It is a Promise 15200.
I did not find any option in the iSCSI device to allow only 1 connection.
--
--
Steven
May you have the peace and freedom that come from abandoning all hope of
having a better past.
--- - --- - - - - - - - -- - - - --- - ------ -
- --- - - -- - - - -- - - -
"Rodney R. Fournier [MVP]" <rod@die.spam.die.nw-america.com> wrote in
message news:%23MiJw%23GuFHA.2064@TK2MSFTNGP09.phx.gbl...
|