Disk Numbering Changes - Reboot required ?

[VMS Support]

One for the panel.

Example Case.
I have a shadow set as follows:

DSA1 = DGA100 + DGA300 + DGA400
The disk are HSG80 based disk.
I need to change the disk id for DGA300.
From DGA300 to DGA800.

1) Connect to OpenVMS (sign into System account).
$DISMOUNT $1$DGA400:/CLUSTER
Check removed from shadow set
$SHOW DEV DSA1

2) On HSG80
Find disk id (In this case say D100).
HSG>SHOW D100
HSG>SET D100 ID=800
HSG>SHOW D100

3) On OpenVMS.
From each node in cluster.
$MC SYSMAN IO AUTO
$SHOW DEV DGA800

4) Add back into shadow set ...
$MOUNT DSA1:/SYS/CLUSTER/CONFIRM/SHAD=$1$DGA800: label

....edit script files that reference disk..mountdisk etc..etc...

Will OpenVMS configure disk with a new id after the sysman io auto.
The disk already has a data structure under the old ID (DGA400). Will I get new one for the disk known as DGA800 ?

Or is it reboot time ?

[Uwe Zessin]

I would do a reboot to create a fresh I/O database in memory - there is more going on behind the scene than just management of the unit identifier.

[Jan van den Ende · Accepted solution]

Well,

I think a reboot sometime convenient would be cleaner, but AFAIK no need to hurry.
And the reboot may be executed as a 'rolling reboot'.
The ONE thing you definitely have to make REAL sure of, is NOT to re-use the old ID (DGA400) for another drive until ALL nodes of the cluster have rebooted!

Proost.

Have one on me.

jpe

[Andy Bustamante]

After you change the ID for each node in the cluster you can delete the old path and reconfigure the new device.

$ MCR SYSMAN IO SCSI_PATH_VERIFY
$ MCR SYSMAN IO AUTOCONFIGURE

No need for a reboot, after all this is VMS.

Andy

[Mike Reznak]

Hi,

you can set the old devices

$SET DEVICE/NOERROR/NOAVAIL

to disable error loging and availability.

Mike