[Mailman-Users] The "arch" command and memory usage

[Paul Davison]

Dear All

I am in the middle of converting various mailing lists that we run here 
from majordomo to mailman, and I am really pleased and impressed with 
mailman so far. Well done to everyone involved.

I have one mailing list that I've been running since 1990 (no, seriously, I 
have!). I have some archived digests from 1995 onwards and an older "mbox" 
file with postings from before that, when the list used to run from older 
software (first MMDF, then, later, PP). Am I showing my age yet? :-)

I wanted to convert all the historical stuff into a proper mailman archive. 
I created a wacking great big stormcock.mbox file (this is the name of the 
list -- it's for fans of Roy Harper's music). I ran the "arch" program. 
Some time later my Linux machine complained about running out of VM then it 
hung.

Investigation shows that the arch program appears to read everything into 
memory before spitting out the HTML etc. This is obviously a problem for my 
large archive running on a fairly modest Linux machine (64Meg Ram, 160Meg swap)

Can anyone suggest a better way for me to do it? I note that there appears 
to be a patch to use htdig on the sourceforge.net site. If that's 
compatible with the release 2.0 mailman, I might try that. Any thoughts?

Cheers

Paul.

-- 
Paul Davison, Managing Director   | Tel: +44 20 7267 2561
SRL Data, 1 Perren Street         | Fax: +44 20 7267 1497
London NW5 3ED, UK                | http://www.steamradio.com