mkfs question

[dictum9]

I need to create one of these super large files filesystems, and cannot remember the option.

I know it's something like mkfs -o largefiles but cannot confirm from the (Useless) man page.

So, where is the list of all these options to go with mkfs???They've hidden it well.

[dictum9]

another question, how I minimize the filesystem overhead?

[Victor Fridyev · Accepted solution]

Hi,

newfs -F vxfs /dev/vgxx/rlvolZ
fsadm -o largefiles dev/vgxx/rlvolZ


It seems to me. you can use
newfs -F vxfs -o largefiles /dev/vgxx/rlvolZ
but I usually do this in two commands

HTH

[A. Clay Stephenson]

If you find the man pages useless then you are probably in the wrong profession and certainly the wrong OS. It's much easier if you can move a mouse, enter a few number, and click and create a filesystem. In any event, a "man newfs_vxfs" will tell you the options you need and it is "-o largefiles". In general, newfs is easier to use than mkfs because it is simply a front-end for mkfs. Of course, if you had bothered to look under the "SEE ALSO" section of the useless mkfs man pages you would have found this yourself.

[A. Clay Stephenson]

how I minimize the filesystem overhead -- with respect to what? You question is much too ambiguous to answer. For example, if you want to minimize the "overhead" of the UNIX buffer cache then you use the vxfs mount options -o convosync=direct,mincache=direct. Of course if you want to minimize buffer cache overhead then an even better answer is to use raw i/o and then no matter how ambigious your question about filesystems is, raw i/o is the answer because it's difficult to have less than zero "overhead".

[Steven E. Protter]

Shalom,

newfs is a shell or wrapper that issues mkfs comands.

From experience this will definitely work. I avoid working with mkfs whenever possible.

SEP

[dictum9]

they syntax for making a filesystem seems to be the same.

[Atul Gautam]

Hi....

To create a super large files filesystem, you can use the following command ----

===========================================
# mke2fs -c -L -O has_journal -T largefile -V /dev/
===========================================

and in place of largefile, you can also try to use largefile4 option with -T.

Here,
largefile --> has one inode per megabyte
largefile4--> has one inode per 4 megabytes



Regarding minimizing the filesystem overhead, you can probably use the "tune2fs" command and the options with that which you can see by going through its man page.




Cheers!


ATUL