dave.toal@t-t.com (Dave Toal) writes: [...]
Just a w.a.g., but (vol in k) / 4 = maxfiles, I think.
Hmmm, though. The numbers don't quite match.
Ah, but they do if you massage them right! :-)
See the thread "Quick question" (sic!) in mid-August 2000 for more details.
statler> df -i /vol/spool Filesystem iused ifree %iused Mounted on /vol/spool/ 190362 4809618 4% /vol/spool/ statler> df /vol/spool Filesystem kbytes used avail capacity Mounted on /vol/spool/ 31776308 6277576 25498732 20% /vol/spool/ /vol/spool/.snapshot 0 5920 0 ---% /vol/spool/.snapshot statler> Connection closed by foreign host. bash-2.02# expr 31776308 / 4 7944077 bash-2.02# expr 190362 + 4809618 4999980 bash-2.02#
So spool is not at max inodes.
statler> maxfiles spool Volume spool: maximum number of files is currently 4999980 (190363 used). statler> maxfiles spool 999999999 Max inode count cannot exceed 8385415 statler>
bash-2.02# expr 8385415 * 4 33541660 bash-2.02#
not 31776308.
The size of the volume is "really" 35307008 KB [34500-20.5 MB: one right-sized "36 GB" data disk?] but only 0.9*35307008 = 31776308 KB can be used, and that's what "df" shows.
The maximum number of inodes allowed is then 35307008/4 = 8826752, but the inode table is not allowed to become more than 95% full, so this corresponds to a maxfiles value of 0.95*8826752 = 8385415 [well, if you round up, anyway].
This same playing around with factors of 0.9 and 0.95 is necessary when computing the minimum (and default) maxfiles value, using 32 KB per inode rather than 4 KB per inode.
Chris Thompson University of Cambridge Computing Service, Email: cet1@ucs.cam.ac.uk New Museums Site, Cambridge CB2 3QG, Phone: +44 1223 334715 United Kingdom.