How to determine if an vxfs filesystem was created with the largefiles option?

[Ralph Grothe]

Hello,

from the manpage mount_vxfs I can only find the not overly helpful hint that if one tries to mount an vxfs filesystem with the largefiles option, and it works then the filesystem can accomodate largefiles.

See, I cannot temporarily unmount the suspicious filesystem to undergo the test,
thus I need another way to determine if the filesystem was created with that option.

Maybe the fstyp reveals this?

Can anyone see this feature from this dump?

# fstyp -v /dev/vgreorg/lvol1
vxfs
version: 3
f_bsize: 8192
f_frsize: 8192
f_blocks: 9864192
f_bfree: 4122779
f_bavail: 4090570
f_files: 1030720
f_ffree: 1030688
f_favail: 1030688
f_fsid: 1074724865
f_basetype: vxfs
f_namemax: 254
f_magic: a501fcf5
f_featurebits: 0
f_flag: 16
f_fsindex: 6
f_size: 9864192

[monasingh_1 · Accepted solution]

fsadm -F vxfs /mountpoint

or use prealloc to create a file larger than 2 GB. (man prealloc)

[Claudiu Schmidt]

Hi,

f_flag shows you the type:
f_flag: 16 = largefiles
f_flag: 0 = no largefiles

Rgds,

Claudiu

[monasingh_1]

fsadm -F vxfs /mountpoint

or use prealloc to create a file larger than 2 GB. (man prealloc). If it fails , no largefiles, if does not then yes..

[Sudhakar J_1]

Hi,

you can use mkfs -m special-file command to see if the large files option is used to create the fs.

Regards

Sudhakar J

[Christian Gebhardt]

Hi

f_flag: 0 nolargefiles
f_flag: 16 largefiles

Chris

[John Palmer]

Try:

mkfs -F vxfs -m /dev/vgreorg/lvol1

This will display the command used to create the filesystem.

As to the fstyp output, from experiment, the value of f_flag is 16 if largefiles is set, 0 if not.

Regards,
John

[Gbenu]

Hi,

You can also see the attributes of your file system in the /etc/fstab file. vi the file and you shold see...

The fourth column of the /etc/fstab file will have an entry like no-largefiles when large file is not allowed and largefiles when it is allowed.

Gbenu

[Ralph Grothe]

Thanks folks,

I really should have given a little more thought before posting.

I simply created a /dev/zero

and dd'ed a 3 Gig file from it into the filesystem.

Since this worked without any problem I for sure know that the filesystem is largefiles capable.

Anyway, your suggestions are also valuable.

Rgds

Raph

[Juan Manuel López]

If you have got JFS online installed, use fsadm .f vxfs /mount_point to get the caracteristics of your fs.

If not, you can take a look to /etc/fstab, there is an indication to the system to mount the fs with their properties.

Other wise, use fstyp -v like other answers tell you.

I hope this help you.

Juanma.

[Frank Slootweg]

A simplification of the other answers (the others were *close*):

fsadm /dev/vg??/rlvol?

Works for both JFS/vxfs and HFS and does not need OnLine/Advanced JFS.

I.e. for example:

# bdf /opt
Filesystem kbytes used avail %used Mounted on
/dev/vg01/lvol3 2048000 1773829 265921 87% /opt
# fsadm /dev/vg01/rlvol3
largefiles
#

[Seth Parker]

I may be missing something, but can't you just go into SAM and look at "View More Information..." on a given filesystem? "Allow Large Files" is one of the things shown.

Since SAM has to get this info from someplace, there's obviously a command, but when you can't remember what it is, wouldn't SAM be a good fall-back?

Seth

[Ralph Grothe]

Addendum Question,

is it posssible to apply the largefile option a posteriori on a filesystem like such
(provided we got OnlineJFS)

fsadm -F vxfs -o largefiles /some/mountpoint

at least the manpage of fsadm suggests this.

[John Poff]

Hi,

Yes, with OnlineJFS you can change the mount options on the fly. We've done plenty of it here. Cool stuff!

JP