output "bdf" isn't correct

[centurion]

/dev/vgsap/lvsapdata5
524156928 51168 520025208 0% /oracle/SID/sapdata5

There is nothing stored in the filesystem. But there is a lack about 4GB ?!

HPUX 11.31

[Jose Mosquera]

Hi,

Hints:
Before mounting the file system ensures that mounting point directory are empty.
If you have used some .dbf file inside of this mounting point you must shutdown your Oracle instance to release used space.

Rgds.

[Manix]

Looks like a very big file system 524 Gigs
and there are some meta data space reservations
for every vxfs file systems say around 1% or slightly less than that.

However in your case it`s < 1 % ,4 Gigs.

This is pretty much expected behavior.

Thanks
Manix

[centurion]

oracle is not running at the moment. there are only a few folders without content.

at the production system I have the same size of filesystem, but there it's ok:

/dev/sap/lvsapdata1
524156928 523573856 578568 100% /oracle/SID/sapdata1


I remount all filesystems, but no way.

Any ideas?

[Jose Mosquera]

Hi,

Please check if your file system is really free or is busy by any process:
#fuser -u /dev/vgsap/lvsapdata5

Rgds.

[Jose Mosquera]

Hi,

Have you checked the current file system consistency by "fsck" command?

Rgds.

[centurion]

The filesystem is free..

root@de06s209 [/oracle/TF3]# fuser -u /dev/vgtf3/lvsapdata5
/dev/vgtf3/lvsapdata5:

[centurion]

fsck is working without errors

[Jose Mosquera · Accepted solution]

Hi,

How many space report "du -k" command?

Also please remember that your disk manager subsystem reserves several space for internal uses, metadata, etc.

Rgds.

[centurion]

the du -sk looks good:

8 /oracle/SID/sapdata5


About reservation/metadatas, see my entry above about our production system. There it's ok.

[Jose Mosquera]

Hi Centurion,

I agree with the previous assessment of Manix, please consider spending of 1% or less for internal management of your disk manager.

Rgds.

[centurion]

Jose, I'am with you. But are you shure?

Because on our other system, the same configuration, we haven't these less for internal management ?!

[Jose Mosquera]

Hi again,

Have you the same patch level in both systems?

[centurion]

the sole difference is, that the affected system is a vpar..

[Jose Mosquera]

What show you this output command in both systems:
#fstyp -v /dev/vgsap/lvsapdata5

[centurion]

system with bad bdf:

vxfs
version: 6
f_bsize: 8192
f_frsize: 8192
f_blocks: 65519616
f_bfree: 65513220
f_bavail: 65001398
f_files: 16434400
f_ffree: 16378304
f_favail: 16378304
f_fsid: 1073872907
f_basetype: vxfs
f_namemax: 254
f_magic: a501fcf5
f_featurebits: 0
f_flag: 16
f_fsindex: 9
f_size: 65519616



system with good bdf:

vxfs
version: 6
f_bsize: 8192
f_frsize: 8192
f_blocks: 65519616
f_bfree: 9194609
f_bavail: 9122777
f_files: 2308640
f_ffree: 2298624
f_favail: 2298624
f_fsid: 1073807370
f_basetype: vxfs
f_namemax: 254
f_magic: a501fcf5
f_featurebits: 0
f_flag: 16
f_fsindex: 9
f_size: 65519616

[Jose Mosquera]

Are you using a Disk Array, is the same for both? Or are using separate disk devices?
Same RAID Layout for both?
Are there the same mount options for both boxes on /etc/fstab files?

Have you created both Volume Groups with the same physical_extent_size, please attach the outputs of:
#vgdisplay -v vgsap

[centurion]

everything is the same (storage-box, raid, pe, pe-size,..).

I made now a newfs and I have the same value as before. I think there is a difference to the hpux vpar-vesion. :-(

[Torsten.]

I'm 100% sure it is not related to vPars...