Dan Nelson wrote: > In the last episode (Jun 08), Kelly Jones said: >> What UFS-like filesystem has unlimited inodes, but is a drop-in >> replacement for ext3, and is fairly easy to configure? >> >> Is UFS2 no longer considered the "best" general-use filesystem? >> >> Reason I ask: I'm going to create many small (~1K) files on a 100G >> disk and thus need at least 100M inodes. >> >> "newfs -i" maxes out at ~52M inodes (862 groups * 60864 inodes =~ 52M >> inodes): >> >> # newfs -N -i 1 /dev/da1;: same results as -i 2048 >> >> /dev/da1: 102400.0MB (209715200 sectors) block size 16384, fragment size >> 2048 using 862 cylinder groups of 118.88MB, 7608 blks, 60864 >> inodes. >> >> I realize I can use "f 512 -b 4096" to get 200M+ inodes, but I'm willing >> to experiment w/ a new filesystem, provided it behaves mostly like UFS. >> Thoughts? > > At this point you're sort of out of the general-use category :) You want > ZFS. Or rather, you don't want to try and fsck a UFS filesystem with 200M > inodes. The three drawbacks I can think of to ZFS are it's hard to boot > from (although you probably aren't booting from da1), it requires more > memory than UFS, and there is no ACL support at the moment (not that many > people used the ACL support for UFS). If you're already on an amd64 system > with 4GB or more RAM you'll be fine. >
Or store your data in a RDBMS rather than in the filesystem. Cheers, Matthew -- Dr Matthew J Seaman MA, D.Phil. Flat 3 7 Priory Courtyard PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate Kent, CT11 9PW, UK
signature.asc
Description: OpenPGP digital signature