Operating System - OpenVMS
1828584 Members
2679 Online
109982 Solutions
New Discussion

OpenVMS Standalone Backup versus Image Backup.

 
SOLVED
Go to solution
Ian R Norris
Occasional Contributor

OpenVMS Standalone Backup versus Image Backup.

If you have two nodes A and B in a cluster each with there own system disk and you do an image backup of system disk A from node B so that node B has no open files on system disk A, is that equivalent to doing a standalone backup of disk A?

I'm just trying to find a way of doing fewer standalone backups.
3 REPLIES 3
Jim_McKinney
Honored Contributor

Re: OpenVMS Standalone Backup versus Image Backup.

Assuming that node B has no open files on node A's system disk and if node A is shutdown while node B is used to backup this system disk, then the backup is equivalent to a standalone backup.
Galen Tackett
Valued Contributor

Re: OpenVMS Standalone Backup versus Image Backup.

If node A is down, and node B has a direct path (i.e. isn't dependent on an MSCP served path) to node A's system disk, then using node B to back up node A's disk should produce as safe a backup as standalone.

If node A is up, then _it_ would likely still have files open for write on its system disk. If that's true then a backup made by node B would NOT be as safe as a standalone backup.

This is true whenever a disk as files open for write, no matter what node has opened the files. There might be other risks as well.

(And of course, if node B doesn't have a direct path/depends on MSCP for access to node A's disk, then you can't back it up at all if node A is down.)
Richard Whalen
Honored Contributor
Solution

Re: OpenVMS Standalone Backup versus Image Backup.

If you mean use node B to do an image backup of node A's system disk while node A is running, then no, this is not the equivalent of doing a standalone backup of node A's system disk.

Even though node B doesn't have any files open on node A'a disk, node A does. The process doing backup on node B will detect any that are open for write when it does the backup because it won't be able to get exclusive read to them.

The question then becomes, are there any files open for write on node A's system disk that are extremely valuable. If you have moved the authorization and queue files to a common disk for this cluster (so that you can continue to operate in case one node goes down), then the files that remain open for write on the system disk are various log files that system processes write to. If you can tolerate loss of some of the data in these files, then you may not need to do standalone backups.