UFSDUMP OF SEVERAL VOLUMES

[inma]

I NEED TO DO A BACKUP OF SERVERAL VOLUMES WITH ONLY ON UFSDUMP IS IT POSSIBLE TO DO IT?

 

[KenCunningham]

Not with only one ufsdump command, but we put the following in a script and run the script to ufsdump the filesystems sequentially:

ufsdump 0fuc /dev/rmt/0cn /
ufsdump 0fuc /dev/rmt/0cn /usr
ufsdump 0fuc /dev/rmt/0cn /export/home
ufsdump 0fuc /dev/rmt/0cn /oracle
ufsdump 0fuc /dev/rmt/0cn /data
ufsdump 0fuc /dev/rmt/0cn /reporting
ufsdump 0fuc /dev/rmt/0cn /index
ufsdump 0fuc /dev/rmt/0cn /redologs
ufsdump 0fuc /dev/rmt/0cn /bull

presumably the same approach will work for you? Note that 0cn is the compressed no-rewind tape device, yours might be 1cn or whatever. Post back if you need more details.

 

[inma]

Thanks KenCunningham, I want to put it in the crontab
The question is that these volumes are great and when run the second ufsdump, a message says that device is busy.

 

[KenCunningham]

Ours is run from cron. Perhaps you aren't closing down the application being dumped? In our case (obviously) it's an Oracle application, and this is shut down before the dump is taken, thus there is no chance that the device on which the Oracle database resides might be busy. Is this the case with you? If not, can you post the exact error message you're getting? Perhaps you could also post the script (suitably amended for privacy) you're using.

 

[inma]

I need to backup many *.jpg files of some directory.
This is the script:
(
ufsdump 0uf dev/rmt/1n /vol13/MT00001A
ufsdump 0uf dev/rmt/1n /vol12/MT00002A
ufsdump 0uf dev/rmt/1n /vol12/MT00003A
) 2>&1 | tee -a backup1.log

When run the second ufsdump, the device is busy until first ufsdump stops, and so this message appears:
DUMP: NEEDS ATTENTION: Cannot open `server02:dev/rmt/1n'.
o you want to retry the open?: ("yes" or "no") DUMP: The E
TIRE dump is aborted.

 

[bfitzmai]

Inma,
You may want to put your dump command through a process like xargs. Make a text file with a list of partitions you want to back up. Then in your script run:

cat <text file> | xargs -l ufsdump 0uf /dev/rmt/1n

Make sure the <text file> entry has the path included.

I think the xargs command will wait until a task complete before it runs the next entry in the list.

 

[KenCunningham]

It sounds (though I can't be sure) that the dumps are trying to run concurrently. I'm not at work at the moment, but I'll check ours when I next get the chance. Otherwise, bfitzmai's suggestion sounds reasonable.