Large directories on Minix partitions.

[Michael Smith <miff@apanix.apana.org.au>]

Here's one that I think is not likely to stress too many people, but
I'd like an educated opinion...

During testing of my Wnews port (which I may be dropping in favor
of another package anyway 8) I encountered a nasty 'feature' of its
inews, whereby a copy of a message is made in the junk newsgroup
for every non-matched newsgroup on the newsgroups: line. 
Nasty. Especially when I was soak-testing it with 11M of compressed
newsbatches.

Anyway, I stopped unbatch after about nine hours and went looking for
the problem.  /var/spool/news/junk was 80K long, an obvious candidate.

Ok, so rm -rf junk

After about an hour and a half, I concluded that the filesystem may well
be faulty, so I rebooted. (at this point, no key response, no vconsole 
switching, nothing. disk is active and light flashing, but no input)

(interesting sidebar : I have 3 minix fs', on two disks : second (63M)
and third (120M) on one disk, all disk (96M) on the other.  I can
fsck the first and second partitions running init/vconsoles, and
I can fsck the second and third running gem/toswin, but the odd one out
gives me 'cannot open device x:' ... weird 8)

Anyway, fsck says the filesystem is okay, so I take another approach :

ls -1 >/tmp/foo

edit foo to look like 

echo XXXXXX; rm XXXXXX

and then nice bash </tmp/foo &

There were about 1100 files left at that point, and it took a good
couple hours to delete them all.  This is on a TT, with an admittedly
rather slow disk (70M MFM running RLL on an Adaptec 4070), but even so,
five seconds per file is a bit sluggish.  This also totally spammed 
the machine, but the progress indicator gave me faith.

Is it meant to be this slow?  If not, maybe there's something useful in
there for future performance enhancement. 8)

Regardless, I'm not planning on having a thousand files in a directory 
in any sort of hurry again 8)

-- 
# mike smith : miff@apanix.apana.org.au - Silicon grease monkey        #
# "The question 'why are the fundamental laws of nature mathematical'  #
# then invites the trivial response 'because we define as fundamental  #
# those laws which are mathematical'". Paul Davies, _The_Mind_of_God_. #