Changing device names on Open VMS server

[Swain]

After restoration of image backup of a different server, our system is showing the device names of the restored system. how to change the device names at next reboot?

e.g., the device names are PHA$DKB100: and PHA$DKB300:
I want to change it to PPA$DKB100: and PPA$DKB300:

[Robert Gezelter]

Amaresh,

Is the goal to move a server, or is the goal to have both the old and new servers online at the same time?

If the goal is to change the name, ip address, etc., it is necessary to change the ip address (in whichever IP stack is in use), and the DECnet address in the DECnet database, as well as the SCSID and SCSNAME in the system parameter file, among a few other things.

What is the goal of this project?

- Bob Gezelter, http://www.rlgsc.com

[Volker Halle]

Amaresh,

you've restored the system disk backup from another OpenVMS server, right ?

You will need to change the system parameters, especially SCSNODENAME. But be aware, that changing the SCS nodename will also need further change, like /ON=xxx:: for the queues. Your TCPIP configuration will also need to be re-done, if you change the SCS nodename of your system.

Volker.

[Swain]

Bob,

I want both the servers to be online simultaneously. The system disk of HPA server has been restored into system disk of PPA.So, PPA is now showing HPA devicenames.I want to change them to PPA.

[Ian Miller.]

The first part of the device name e.g. PHA$ is set by the SCSNODE system parameter.

So it appears you need to change that and the other places that the node name appears.

Do read the information about this in the VMS FAQ
http://64.223.189.234/vmsfaq/vmsfaq_007.html#mgmt9

[Jan van den Ende]

Amaresh,

this is leading to some clash.

The PPA$ and PHA$ is directly taken from the computer name.

But if you make those names the same, you should make REALLY sure thos two system are NEVER EVER up at the same time!

The way around this (since the VMS V4 or V5 times) is to set up ALLOCLASS.

If you define the SYSGEN parameter ALLOCLASS to a non-zero value (say, for this example, 123) then the device names will become $1$123DK.

Now all you need to do is make sure thet are not part of one and the same cluster. (the mechanism was introduced mainly to differentiate between the same devices WITHIN a cluster, by choosing different ALLOCLASS versions, but it is also ideally suited for your purpose)

By the way, another way around your problem, (and more in line with "the VMS way of doing things") is to define SYSTEMwide Logical Names (LNMs) for the devices, once during bootstrapping, and further on ONLY reference the device by there LNM. That way you create independency from the hardware names.

If you should (have to) change hardware sometime in the future, eg, move to SAN device, then your hardware names WILL change (to DG for SAN disks).

By using LNMs, that "problem" is also solved.

Success!

Proost.

Have one on me.

jpe

[Swain]

Jan,
I think you have suggested this for cluster system. In my case, the servers are not in cluster. Thanks for the info.

[Swain]

thanks Bob,Ian,Jan,Volker
Changing the node name will do (refer the below link)
http://64.223.189.234/vmsfaq/vmsfaq_007.html#mgmt9

[Jan van den Ende]

Amaresh,

>>>
Jan,
I think you have suggested this for cluster system. In my case, the servers are not in cluster. Thanks for the info.
<<<

So please, carefully read again.

The $ device names are _VMS_, and NOT cluster-specific.

The same goes for ALLOCLASS and $$ device names.

They work also (with the effects you seem to desire) in standalone systems.

And the device-type dependent hardware names are at the core of DEC systems since before VMS 1.0. _IF_ you move to different drive technology, you WILL move to different device names. Period.
Therefore, the LNM mechanism was originally introduced: de-couple hardware naming from reference naming.

hth

Proost.

Have one on me.

jpe

[Swain]

Jan, Thanks a lot.