Re: determining the bytes-per-inode (mke2fs) value

joysons3@xxxxxxxxx staggered into the Black Sun and said:
I have a running disk array. The data set is basically huge from the
radio telescope, but some others are small (descriptions). Now there
will be a new disk array (and [I'll] copy the current one to the new
[one]) and I am formatting this new one [with] mke2fs. I was told that
[the] ext3 filesystem can be [fine-tuned] at the bytes-per-inode value
according to the real file size situations, to be 4k, 8k, 1M,
4M, etc.

Is there a benchmark method that I can run it on my current running
disk array, and report/suggest [what] the bytes-per-inode [should] be?

df -i (tells you how many inodes are being used)
df (tells you how many 1K blocks are being used)

....then do the obvious thing. Remember that the number of inodes is set
at filesystem creation time for ext23, and can't be changed without
remaking the FS. So make more inodes than you think you need, or you'll
lose spectacularly. If most of the files are > 1M, then it's usually
safe to use -T largefile. HTH,

--
"Oh bother," said the Borg, "We've assimilated Pooh."
--MHR on alt.fan.pratchett
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
.