Error : sys.unix.fs.advfs.fset.backup.unlock

[so.nimda]

Hi all,

Noticed this error in my vdump log :

vdump: Status at Fri Aug 4 23:30:28 2006
vdump: Dumped 4833557141 of 4833826811 bytes; 100.0% completed
vdump: Dumped 618 of 618 directories; 100.0% completed
vdump: Dumped 22450 of 22450 files; 100.0% completed
vdump: Dump completed at Fri Aug 4 23:30:28 2006
vdump: error [34] posting event: sys.unix.fs.advfs.fset.backup.unlock
The name of the modified device is /dev/ntape/tape0
path : /oracle/stage/817_64
dev/fset : sap_dmn#oracle_817
type : advfs
advfs id : 0x3c731801.000809af.12
vdump: error [34] posting event: sys.unix.fs.advfs.fset.backup.lock
vdump: Date of last level 0 dump: the start of the epoch
vdump: Dumping directories
vdump: Dumping 1048117637 bytes, 1733 directories, 3161 files
vdump: Dumping regular files

Has anyone encountered this error before? Also, I've noticed that /usr directory has reached 100%. could it be due to the advfs error?

Any help is appreciated.

Thanks & regards

[Mark Poeschl_2]

I'm guessing it's more a cause than a symptom. Do you have /var installed under /usr? If so and you're not doing enough clean-up, the EVM facility won't be able to log events any more - hence the error you're seeing. Just a shot in the dark...

[Victor Semaska_3]

rancid,

I usually see this error when I do vdumps in single-user mode. It's because the evmd daemon isn't running. So I think Mark is correct on this.

Do you see evmd running?

Vic

[so.nimda]

Hi Mark & Victor,

Thanks for your reply.

I have rebooted the server and did the vdump again and the error isn't there anymore.

Unfotunately, I didn't check if evm daemon was running before the reboot. When I check now, I see that it is running. :(

Also the problem with the /usr & /var reaching 100%, any ideas about that?

Both these directory is in it's own domain and I'm not able to find what is causing the growth. I found some core files and I've deleted them but still the percentage has not gone down.

I'm thinking some log files are growing but I'm not able to find them :(

Regards

[Victor Semaska_3]

rancid,

If quotas are enabled (userquota,groupquota in /etc/fstab), I'd 1st make sure the quota database isn't corrupt with the command:

# quotacheck -av

If that doesn't help I'd then look for large files with the command:

# find /usr -xdev -size +1000k -exec ls -s {} \; | sort -k1,1nr | more

This will find files 1MB or larger and list the largest ones 1st. You can adjust the size of files to look for with the -size option.

Finally I'd look for directories with lots of small files using lots of space.

# find /var -xdev -type d -exec du -sk {} \; | sort -k1,1nr | more

Note the sizes are in KBs. Also, if those disks have lots of files in them then the commands will take awhile to run.

Vic