Fbackup on HPUX 11:00

[Tim Skelhon]

I am running fbackup within a backup script that shuts and restarts oracle
before and after the fbackup.
If the fbackup gives acceptable (to me) WARNINGS (eg: file not backed up) it
exits with code 4.My script tests for a 0 as success and restarts oracle if
fbackup returns a 0.If I test for 0 and 4 and treat them both as successful I
can get round this problem.
However,there are other things that cause fbackup to exit with a 4 (eg: run out
of space on tape) and I want to treat these as failures - How can I get around
this?
Does anyone run fbackup within a script and test for success but do it in a
different way?
This problem does not happen at 10:20.

[Sachin Patel_2]

Hi
I run fbackup from script. I used
fbackup ???????? | /bin/tee -a logfile
then set RESULT = `tail -1 $LOGFILE | awk '{print $4}'`
set RESULT2 = `tail -5 $LOGFILE | grep backup /
did not complete" | wc -l`
from value of RESULT2 I determine the sucess of my backup.

[Gian antonio D'addetta_3]

Hi,

i have some what similar problems. Since we installed the new Y2K patches on
our HP-UX 10.20 machines i get mostly '4' as exit code, although it seems that
the warnings are acceptable. Before ipdating the system i never had this
problems, the exit code was always 0, even if there have been some warnings.

If the tape space is not enough, usually i had the exit code '2' and a query,
that i should enter a new tape, so i guess that there shouldn't result a
problem with the available tape space and teh exit code '4'.

I tested the fbackup through the br_backup scipt and found that since the Y2K
updates the sensibility to warnings of the fbackup command got higher and so
results in exit code '4', although almost very few warnings (like some active
files during backup) apperaed. I'm still searching for the answer.

Bye

[Boon ping Teo]

I have also encountered exit code 4 problem. There is no problem while listing
the files that were backup on the tape. It takes up to about 8GB and the tape
is able to record up to 12 GB.
Anybody can help ?

[Vince Arends]

Have you tried grepping the fbackup log file for the specific error? If you get return code 4, then grep the log file for specific text to determine if you want to continue or abort.

[Antoanetta Naghiu]

Tim,
I while ago, after upgrate from 10.20 to 11.0 I had the same issue. A call had been placed with HP at that time. They recommended to install a patch but did not help. Following up with them, here you go HP's work around:
Sam uses the following script to back the backup :
/usr/sam/lbin/br_backup
The line that would want to replace is line 329 or if [ ?$exit_code? = ?0? ] with if [ ?$exit_code? = ?0? -o ?$exit_code? = ?4? ]
The section that the line looks like the following :


#
#
# Save the exit code and write a completion message to the logfile and to # the specified mail recipient.
#
if [ ?$exit_code? = ?0? ]
then
if [ ?$mailee? = ?? ]
then
echo ?nl_lookup br_scr 8 ?%s backup completed on %s: %s.? -v
$text -c ho
stname -c date? >> $log_file

The SR number for correcting this problem is 1653263095
--

[James R. Ferguson]

Tim:

My personal preference in similar situations is to run fbackup in verbose mode (-v) to capture status messages into a log file; to grep for "key" messages; to establish one's own acceptable tolerance for the presence or absence of the "key" messages; and mail an operations account with the status of the backup, AND the messages.

I found that relying solely on exit values didn't provide me with the level of comfort that I wanted.

...JRF...