fbackup (9999): Interrupted system call

[Luc Monnens]

Hi all,

In the middle of my fbackup log I got:

fbackup (9999): Interrupted system call
fbackup (9999): Interrupted system call
fbackup (2004) WARNING: reader 1: 1st read error, fbackup (2004): WARNING: reader 0: 1st read error, firstread=32768, n=31744
firstread=32768, n=0
fbackup (9999): Interrupted system call
fbackup (2004) WARNING: reader 1: 1st read error, firstread=32768, n=0
fbackup (9999): Interrupted system call
fbackup (2004) WARNING: reader 1: 1st read error, firstread=32768, n=0
fbackup (9999): Interrupted system call

and so on and on and on repeating the last two lines and then it continues as normal and finally finished normally too.

What are these interrupted system calls?
What is causing them?
And most importantly what is it doing to my backup? Can I still rely on it to have a backup of all my files?

Any input is much appreciated.

Kind regards,


Luc Monnens

[dongming]

you cann't use that backup files.

the problem maybe relate many problem.
you can follow this way.
1. first to change tape dirve is ok
. try to tar a small file to drive
. try to fbackup a small file to drive.
if this error message display again. try to replace new media or clean the drive.

2. if it is ok on first step.
try to fbackup again. if the problem still exist. you should find some patch relate fbackup and download to install it.
then try again.

[Jeff Schussele]

Hi Luc,

Check your /var/adm/syslog/syslog.log as I suspect you may be having a HW (possibly disk?) error or bumping up against some sort of system limit . If so the syslog should give you a clue.

HTH,
Jeff

[Jino.P.V]

Hi Luc,

I think your disk is having physical damage(Bad block). Check for that. If it is physical damage fsck will not recover it.

regds,

Jino

[Mahesh Kumar Malik]

Hi Luc

1. Check /var/adm/syslog/syslog.log file for errors

2. It could be a problem with tape drive or medis. Clean the drive head and use another good media

Regards
Mahesh

[Luc Monnens]

Thank you for the input.

Some more info:
1. After these warnings the backup continuous and completes successfully.
2. We use another tape every day.
3. There is nothing in the syslog.
4. I mount these disks to the twin server in an HA cluster and run the same backup without any warnings there.
5. The firstread=32768 is there a few hundred times every day next to a couple other numbers.

Is there a specific meaning to this number "32768"?
What are these "interrupted system calls"?

Kind regards,

Lu

[Marlou Everson]

For some reason, the reader process is having a problem. Perhaps a memory problem or kernel parameter? You may want to check if you have the latest fbackup patches.

The backup is probably okay but may not have written all the files to tape. If you are creating an index (-I option) then look at it for any entries beginning with XXXX. These are the files that were not written to tape. You can always verify the tape with frecover -x -vN.

Marlou

[Luc Monnens]

Maybe the problem was due to accessing the same files from 2 servers simultaneously.