OpenVMS IA64 Version V8.3 restore

[Dax Yap]

Hi,

I have done a backup to a internal disk with the command: backup /log DKA200:[*...]*.*;* []dka200.bck/sav

and move to DKA100:[TEMP]

When finished, the directory/files structure is good:
backup /list dka200.bck/sav

However, when I restore to another disk with the command BACKUP/LOG dka200.bck/SAV DKA200:

All the files restored to DKA200:[TEMP] as flat files without directory structure.

Please help me out. Appreciated it.

Thanks

[John Abbott_2]

Hi, if you want to backup and restore a complete volume, then please see the online help and/or documentation on the BACKUP /IMAGE qualifer.

The above command does correctly from what you describe (i.e. into a single directory) as that's what you asked it to do!

Some suggestions;

Save operation
$ backup/image source: target:
$ backup/image source: target:saveset.bck;/sav

Restore
$ backup/image source: target:
$ backup/image source:saveset.bck;/sav target:

You might want to consider initializing the target volume beforehand (see init and qualifers /headers /window /limit /index /struct /clust) and backup with /noinit on the restore.

Regards
John.

[Dax Yap]

Hi John,

Thanks for the suggestion.

I have tri-node clustered. Can I save the image to another node disk? How-to?

The problem is I want to restore to a server which a distance a way. Thus, I will ftp the huge image across.

Thanks for your help.

Cheers.

[Jan van den Ende · Accepted solution]

Dax,

To begin with: WELCOME to the VMS forum!

>>>
backup /log DKA200:[*...]*.*;* []dka200.bck/sav
<<<

and

>>>
BACKUP/LOG dka200.bck/SAV DKA200:
<<<

If you want the second operation to be the reverse of the first (ie, restore what you saved) then you should also reverse the input and output filespecs.

That would make the restore
$ BACKUP/LOG dka200.bck/SAV DKA200:[*...]

Also look at John;s answer: for entire disks the /IMAGE is much efficienter. AND, even more important, /IMAGE also "understands" disk structure.
In your commands, any structure info is lost. For most disks it would only cost a lot of extra time, but if this was a SystemDisk (or if aliasses have been used for other reasons) then, _ONLY_ /IMAGE can handle that correctly.

hth

Proost.

Have one on me.

jpe

[John Abbott_2]

Hi Dax,

>> Can I save the image to another node disk? How-to?

If the "other" node is part of the cluster, then can you not mount this volume locally and create the ss ? You may have to enable MSCP ($ MC SYSGEN HELP SYS MSCP - for a brief read up) but I'd read up a bit beforehand and it requires a reboot.

You can also use DECnet (native or DECnet over IP) with the backup utility to effectively push the target save set over the network

Example to try;
$ backup/log login.com; remote_node"account password"::volume:[directory]saveset.bck;/sav

>> The problem is I want to restore to a server which a distance a way. Thus, I will ftp the huge image across.

Not sure how big your disk is ? but it's possible to backup/image over decnet too (as an output save set).

You could look at ZIP/BZIPing your save set before ftp'ing it (see the freeware site; sorry don't have the link handy).

Regards
John.

[John Abbott_2]

Sorry Dax, my internet connection is bad today, here's the freeware site link

http://h71000.www7.hp.com/openvms/freeware/

It's a trade off of cpu vs network - if you can ftp(bin) the save set over the network overnight (and there's no panic) then I'd leave is uncompressed. Try a small save set and time it. Compressing eats cpus and you've got to unpick the file attributes afterwards ($ set file/attr=(lrl:save_set_block_size,rfm:fix,rat:none) saveset.bck

Good luck
John.

[Dax Yap]

Thanks John and Jan. The restoration is in progress. This disk is not system disk.