HPUX 11.00 : Volume Group Creation w/8k Block Size, Largefiles Enable

[LG Porter]

I have a L2000 Class server running HPUX 11.0 attached to a EMC array. I need to create a volume group with a 8k block size as well as "largefiles" enabled on each logical volume. The default block size for a new volume group is 4k. What is the correct syntax for creating a volume group with a 8k block size? How is "largefiles" enabled on a LVM? Can the "fstyp" command be used to verify "largefiles" enabled?

[Frederic Sevestre · Accepted solution]

Hello,

The range for the PE size is from 1 to 256 Mo.

To create a fs wiyh largefile option :
newfs -F vxfs -o largefiles /dev/vgxx/rlvolxx

To check :
fsadm -F vxfs /

Regards,
Fr??d??ric

[Pete Randall]

And, for your 8k block size you need to specify the -s option on your vgcreate command.


Pete

[Armin Feller]

create a volume group with a pe_size of 8MB (4MB is default):

# vgcreate -s 8 -n vg_name /dev/dsk/c... /dev/dsk/c...

to verify if the lvol (filesystem) is already configured for largefiles:

# fstyp -v /dev/vgXY/lvolX | grep f_flag
---> 16= largefiels
0 = no largefiles

modify an existing filesystem for largefiles:

# fsadm -F -o largefiles /dev/[vg]/[rlvol]
# mount -F vxfs -o largefiles /dev/[vg]/[lvol]/[mountpoint]

create a new filesystem with largefile option:

# newfs -F vxfs -o largefiles /dev/[vg]/[rlvol]
# mount -F vxfs -o largefiles /dev/[vg]/[lvol]/[mountpoint]

edit /etc/fstab and add the largefile option:

/dev/vgxx/lvolx /[mountpoint] vxfs delaylog,largefiles 0 2

Regards ...
Armin

[John Palmer]

Hi,

You are confusing volume groups and filesystems.

Volume groups have a Physical Extent size (PE), the default is 4Mb.
Filesystems have a blocksize (default 1Kb) and largefiles capability.

To create a filesystem with 8Kb blocksize and largefiles, use:
newfs -F vxfs -o largefiles -b 8192 /dev/vg??/rlvol?

Where /dev/vg??/lvol? is the name of a logical volume.

Regards,
John

[monasingh_1]

first of all Volume group or Lvol has nothing to do with your. Block size is a property of filesytem you create on any Logical volume.

to create largefile enabled filesystem, you can newfs the lv with -o largefiles option.

for 8k block size you also give -b 8192 option in newfs with -o largefiles option.

To check lvol if it has largefiles enabled or not, just run
fsadm -F vxfs /mountpoint
where that lv is mounted. it will tell if the filesystem supports largefiles or not.
fstype -v /dev/VGXX/lvolX will show the block size , see f_bsize parameter there.

And your question, how lvm enables largefiles, internally I do not know, in case of vxfs, i think it may be allowing unlimited extents or in HFS case it may be allowing many redirection for block allocation for a file, but this is just a guess, I think HP LVM engineer needs to step in here...

hope this helps.

[Sridhar Bhaskarla]

Hi,

As pointed by others, you got confused with block size and the extent size. To specify extent size you would use -s option with vgcreate. -b with newfs if you intended block size. I would not suggest deviating from the default block size (f_frsize) as there would be no gain plus will consume extra overhead on the space as the minimum block it will reserve would be 8k.

An interesting way of checking if a file system is enabled for largefiles or not use mkfs ( I am serious) command with -m option.

mkfs -m /dev/vg01/lvol1

It will printout the line that contains the options of previous mkfs commands and will show you if largefiles option was previously used.

-Sri