Largefiles - HP-UX 11.23

[Fred Sanford III]

Howdy folks,

I am relatively new to hp-ux and am looking for a definitive answer regarding largefiles on hp-ux 11.23.

I've found some sources of information that state that largefiles are enabled by default when creating/mounting vxfs file systems in 11.23; there is no need to specify when creating, or moutning or when making an entry in fstab. But I also find information to the contrary.

Can one of you gurus out there please tell me which it is?

Thanks in advance!

[Tim Nelson]

Not by default.

I have always had to specify. Might be able to edit /etc/defaults ..

from the man page newfs_vxfs
-o largefiles | nolargefiles
Valid only for the Version 4 and 5 disk layouts.
This option controls the largefiles flag for the
file system. If largefiles is specified, the bit
is set and files two gigabytes or larger can be
created. If nolargefiles is specified, the bit is
cleared and files created on the files system are
limited to less than two gigabytes. The default
is nolargefiles. See mkfs_vxfs(1M),
mount_vxfs(1M) and fsadm_vxfs(1M).

[Sandman!]

One way to find out if an exisiting filesystem has largefiles enabled is to run the command below:

# mkfs -F vxfs -m /dev//

[James R. Ferguson]

Hi Fred:

You may wish to add 'largefiles' or 'nolargefiles' to '/etc/fstab' to match each underlying filesystem. When you do this, mounting will only succeed if the declaration in 'fstab' matches what the actual filesystem's largefiles bit has set.

Regards!

...JRF...

[Fred Sanford III]

Thanks for the input, guys.

The thing is, I have a running 11.23 system with vxfs file systems that were created without the largefile option specicied, and there certainly isn't anything in fstab, yet there are definitely files larger than 2GB on them.

Fsadm shows largefiles, mount -v does not show largefiles, but i assume that is because that option is not specified in fstab?

I guess the safe way to go would just be to proceed as we've done in the past, specify large files when creating the file system, and add the option to fstab. But I hate doing things because "that's the way it's always been done."

[Tim Nelson]

I believe you can just flip the largefiles bit with fsadm. You will still need to umount and remount with the option set in /etc/fstab.

[Dennis Handly]

>I've found some sources of information that state that largefiles are enabled by default when creating/mounting vxfs file systems in 11.23;

I thought so too. But for 11.31 it is the default and mkfs_vxfs(1M) says so.

[Norman_21]

If you have OnlineJFS installed you can simply change the file system attributes online using the fsadm.

First find out if the file system is largefiles enabled. Keep in mind, only file systems over 2GB need to be largefiles enabled.

home/rootlbm > fsadm /tmp
fsadm: /etc/default/fs is used for determining the file system type
nolargefiles

Then set it as following:

home/rootlbm > fsadm -o largefiles /tmp
fsadm: /etc/default/fs is used for determining the file system type

Check it again.

home/rootlbm > fsadm /tmp
fsadm: /etc/default/fs is used for determining the file system type
largefiles

you can also use SAM or update the /etc/fstab as pointed out by my peers!

[whiteknight]

Fred,

For changing the largefiles bit on a JFS filesystem:
To convert an Advanced JFS filesystem:
fsadm -F vxfs -o largefiles /(mountpoint)

To convert a base JFS filesystem (filesystem MUST be unmounted):
#fsadm -F vxfs -o largefiles /dev/vg0#/rlvol#

To mount the new largefiles capable file system :
#mount -F vxfs -o largefiles /dev/vg#/lvol# /(mountpoint)

To set the correct /etc/fstab
syntax :

/dev/vg0#/lvol# /ignite_depot vxfs delaylog,largefiles 0 2

WK

[Roberto Arias]

Hi all:

In version >11.23 yes, is for default the larguefiles option.

you can check if you haver Fs with largue files with fstyp -v , 16 is the same tha largefiles available.

please see man page of fstyp
regards

[Fred Sanford III]

Thanks for al the info!