fbackup hung

[Starrynight]

Hi everybody

I need help!
I have a script that is run everynight, to performe a backup with Omniback II (data) and fbackup (operative system). It follows these steps:
1- shutdown with reboot, 2- mirror split,3- Omniback, 4-fbackup, 5- mirror merge.
It rans for several years without problem but three days from today it hungs in fbackup (and therefore does not make the merge of the mirror).
The output of the comand ps -ef is the following (in the morning):
ps -ef|grep fbackup
root 6742 6739 0 23:58:51 console 0:00 /usr/sbin/fbackuprdr 135177 1 262144 C 0 0
root 6741 6739 0 23:58:51 console 0:00 /usr/sbin/fbackuprdr 135177 0 262144 C 0 0
root 6739 7914 255 23:58:50 console 605:43 /usr/sbin/fbackup -1 -uvg /etc/opt/telecom/BACKUP/graph -f /
root 6743 6739 0 23:58:51 console 0:00 /usr/sbin/fbackupwrtr 135177 262144 4 C 0 0 /dev/rmt/dds1
root 8225 22280 2 10:11:25 ttyp2 0:00 grep fbackup

I have tested the drive and the fbackup comand (with the mirror "merging"). No error found.
Any suggestions?

T?nia

[melvyn burnard]

perhaps the amount of data to be backed up using fbackup has grown such that it sits asking for another tape?
This will be a problem if you are running a script in something like cron.
Do you redirect stdout/stderr to a file? if so what does that say?
just a few thoughts.

[Mark Vollmers]

Hi. I think that Melvin is probably right. You can do a quick check by running something like bdf and see how big the file system is and compare this to the tape size. If you run from cron, root should get an email with any warnings or errors. If you can, try to run it during the day (just the fbackup) from the console. You might have backup errors (since files are being used) but you will see any errors that it gives. If it is the size, take a look for large files that you can remove. Maybe you are right on the edge and a core file snuck in and kicked it over the limit.

Mark

[Starrynight]

The script is not run as a cron job. There?s a menu and someone press the right button. The data to be backed up did not grew up either.
The problem is that, as you can see by the ps -ef comand output, the the fbackuprdr and fbackupwrtr did not started when the father process (fbackup (n? 67739)) launched them.
As I have tested de fbackup comand I can not see where the problem can be.

[Mark Vollmers]

You may have done this already, but have you looked (through vi) at the script that is being run? Make sure that all the commands it calls (like fbackup) are complete. Also, test each sub-part of the script. For example, if you run the big script (backup_script), and it runs part_a, part_b, part_c, and part_d (all seperate commands or scripts), have you tried each individual one? There might be an exit code or something that is preventing the next step from beginning. (backup might finish but it isn't telling the script to move on).
Also, is there any output to the screen when the "button" is pushed to begin? Is this output to a file somewhere? If there is a problem in backup, there should be some messages somewhere.

Mark

[Robert Campbell · Accepted solution]

I had something like this happen to me twice. I run fbackup every night without any problems and then for some reason fbackup would never finish. What I discovered on both occasions was that I had a directory that was created that never ended. So when fbackup was attempting to create the index on the tape it could never get through all the directories. I went back through my incident reports to find how I discovered this. The support center was trying to help me troubleshoot the problem and told me to use du -sk. I received the output from this command that told me a path was too long. I discovered the path was created by a programmer the day that I started having problems. I don't know if this is the same problem as you are having but it could be a possibility.

[Mark Vollmers]

Another thing that I just thought of. Are there any errors showing up in the syslog? I had some problems with fbackup a while ago and it turned out to be linked to SCSI errors that I saw in the syslog; turns out there were some errors on the drive and backup would hiccup when it hit the bad inodes and hang. Might not apply but doesn't hurt to check.

Mark

[Sanjay_6]

Hi Tania,

If it has happened only once, then it seems more likely to be a fluke case of the Device or Cartridge giving a problem. I think i had a similar problem once, next time it worked fine, both the drive as well as the cartridge.

Thanks

p.s. : how did you put that ^ over a in your name ?.

[Starrynight]

Dear Mr. Campbell

I would like to thank you for your kind answer because that was exactly the problem.

For all the others that replayed my help request I am also thankful, and by the way, Mr. Sanjay, the ^ in the a is how my name is written since my mother gave it to me :-)