File System Information

[Bryan D. Quinn]

I am trying to determine the block size that was configured for a file system. I beleive it was created with 8K blocks, but I would like to verify.

Anyone know of a way to determine what the block size is configured to?

Thanks,
-Bryan

[Dave Olker · Accepted solution]

Hi Bryan,

You can use mkfs -m to determine the string used to build the filesystem. For example:

# mkfs -F vxfs -m /dev/lvm01/lvol1
mkfs -F vxfs -o ninode=unlimited,bsize=1024,version=6,inosize=256,logsize=16384,largefiles /dev/lvm01/lvol1

This shows "bsize=1024" meaning a 1K block size.

Regards,

Dave

[James R. Ferguson]

Bryan:

For example:

# fstyp -v /dev/vg00/lvol7

Regards!

...JRF...

[Sanjay_6]

Hi,

did you try "fstyp",

fstyp -v /dev/vg_name/rlv_name

Looks for the field, "f_bsize"

Hope this helps.

regds

[A. Clay Stephenson]

... and I'll bet that you learned in your Oracle class that blocksize is very important. Well, it was in HP-UX land when hfs/ufs filesystems were commonly used but not (when I know you are running vxfs because it is much faster) now. Vxfs filesystems are extent-based rather than block-based so that the filesystem always tries to write and allocate in large chunks. Block size is unimportant except as the smallest quantum of filesize. Leaving your blocksize at 1KiB (the default) does no harm at all from a performance perspective.

[Bryan D. Quinn]

Thanks for everybodies help. I did try fstyp, but like a ding dong did not try -v...duh. Thanks again and a special thanks to Clay for the clarification on the block size. You are exactly right this is an Oracle/SAP implementation and the database people asked me to make sure that I did create the FS with 8192 blocks.

Thanks again for everybodies responses.

-Bryan