fbackup

[Jeanette Radebe]

When running fbackup the system load gets very high. Advised customer to install patches PHCO_21263 and PHCO_21186 , but still gets the same problem were fbackup is using huge amounts of cpu time.

The cu is using the following command:
"nohup fbackup -f /dev/rmt/0m -i / "
Created new device files without compression.
This did not help, still get fbackup running the next mroning and taking up large cpu time.
Still not working. Output for # ps -ef|grep fbackup

root 21827 10644 1 08:07:28 ttyp3 0:00 grep fbackup
root 29612 1 224 23:43:57 ? 272:11 fbackup -f /dev/rmt/0m -i
/
root 29614 29612 41 23:43:58 ? 1:02 /usr/sbin/fbackuprdr 9607 0
2621
44 C 0 0
root 29615 29612 1 23:43:58 ? 1:00 /usr/sbin/fbackuprdr 9607 1
2621
44 C 0 0
root 29616 29612 5 23:43:58 ? 0:52 /usr/sbin/fbackupwrtr 9607
26214
4 4 C 0 0 /dev/rmt/0m

Please advise

Best regards,
Jeanette

[Andy Monks]

I'd suggest using a config file with fbackup.

The good values are :-

blocksperrecord 128
checkpointfreq 256
filesperfsm 1000

Then depending on the tape drive (dds3/4, dlt) you may want to change 'readerprocesses' to 4 or 5.

[CHRIS ANORUO]

In your fbackup config file located at /etc/sam/br/fbackup_config,
update the fields as follows:
blocksperrecord 128
records 32
checkpointfreq 256
readerprocesses 4
You can check the online manual for fbackup to see what the parameters are for.
When doing manual backup, use the (-c with config file) fbackup -f /dev/rmt/0m -c /etc/sam/br/fbackup_config

[Paul Latchem]

It's possible that you may have a corrupt mount point which is causing the fbackup command not to complete. Try verifying that all your filesystems are correctly mounted. Try issuing the ll command in the / directory, if you have a corrupt mount point beneath this directory, then this command will hang and you should then be able to troubleshoot further. Hope this helps.