Seriously? You're worried about the space inodes will take on a TWELVE TERABYTE file system?
What amount of disk space (%) could this actually save? Is it really worth worrying about today? I guess I could see it if it's going to waste >1Gb, if only to save time during the mkfs process... -Adam
Trevor Cordes trevor@tecnopolis.ca wrote:
I'm making (as we speak) a 12TB RAID6 array, using XFS. My first attempt I noticed XFS was making waaay too many inodes. I know in ext3 I can adjust the inode count, like # bytes per inode, etc. Does anyone know how to do the same in XFS? Man xfs is super confusing on inode count manipulation. I have no idea what numbers it wants in its parameters.
Since my fs has mostly large (2-4GB) files, I like to reduce the inode count by 1-3 orders of magnitude from the defaults to give me more usable space.
Thanks! _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
On 2011-05-11 Adam Thompson wrote:
Seriously? You're worried about the space inodes will take on a TWELVE TERABYTE file system?
Yes! :-) I'm so anal I agonized over whether to use the extra 200-400MB over the 2,000,000,000,000 bytes the raw disks provided. In the end I chose to round down the end-sector to the nearest 1MB aligned-sector just under 2,000,000,000,000 bytes; minus 1 sector.
I had an astonishing thought, we're "writing off" 200-400MB like it's nothing, when my very first hard disk was less than half that size! Hmmm...
What amount of disk space (%) could this actually save? Is it really worth worrying about today? I guess I could see it if it's going to
It was apparently making 644 million inodes. I have no idea how big its inodes are, but if we assume a (random guess) of 16 bytes, that's 9.59GB wasted! Ouch. Even my main drives in 1997 weren't that big.
- Getting Old
-
Adam Thompson -
Trevor Cordes