Can I create "largefiles" automatically?

[Stuart Abramson_2]

I keep forgetting to set "largefiles" when I make new filesystems.

I believe that at my last job we didn't have to specify this at creation of every new file system.

Can I put something in /etc/default/fs, or something?

(Now that I have been reminded about this for the 3rd or 4th time, you'd think I'd remember.)

Also, I seem to recall doing this for large files:

newfs -F vxfs -l /dev/vgXX/rlvolNN

buy, my man page says to do:

newfs -F vxfs -o largefiles /dev/vgXX/rlvolNN.

Which is it?

Does this change with different versions of VXFS?

[Vicente Sanchez_3]

Hello,

You can use fsdam -o largefiles /filesystem

HTH, Vicente.

[James R. Ferguson]

Hi Stuart:

You can convert a filesystem to enable largefiles after it is created:

# fsadm -F hfs -o largefiles /dev/vgXX/rlvolN

# fsadm -F vxfs -o largefiles /dev/vgXX/rlvolN

Note the raw device is used.

Regards!

...JRF...

[Stefan Farrelly · Accepted solution]

We always use newfs -F vxfs -o largefiles /dev/vgXX/rlvolNN.

Used it on 11/11i and different versions of JFS (3/3.3/3.5) and it works on all of them.

/etc/default/fs is only used for filesystem type, no options. Only thing I can think of is replace your newfs program with a script which calls the real newfs program but adds on -o largefiles, but next time you install a jfs patch it may replace newfs and your back to square one. How about creating a script called newfs.sh in /usr/local/bin which checks your options and adds on -o largefiles then runs the newfs binary. Then you just need to remember to run newfs.sh when creating new filesystems.

[Sunil Sharma_1]

Hi,

You can not pass this argument by default, it has to be passed with newfs command.

but in Solaris environment you are not required this option for ufs file systems.

Sunil

[Stuart Abramson_2]

Thanks all.

I will just remember to add the "-o largefiles" in my newfs scripts from now on.

Stuart