[Mailman-Users] MHonArc

[J C Lawrence]

On Thu, 3 May 2001 14:39:19 -0500 (CDT) 
Tanya Brethour <brethour at ncsa.uiuc.edu> wrote:

> Quick question: I am converting to use the mhonarc archiver
> instead of the builtin mailman archiver because of the MIME
> attachment issue. I was curious how people integrate the two.

I subscribe a special archiver address to each list I wish to
archive.  

Under that address procmail then files each list into folders.  A
cron job handles renaming the folders each quarter (I group by
quarters rather than by month).  To make procmail happier (and save
editing the procmail RC) I do this by moving symlinks.

  eg
  ~/mail/lists/somelist-l/current -> 2001Q1
  ~/mail/lists/somelist-l/2001Q1
  
  After cronjob runs on April 1st:
  ~/mail/lists/somelist-l/current -> 2001Q2
  ~/mail/lists/somelist-l/2001Q1
  ~/mail/lists/somelist-l/2001Q2

  procmail files into the "current" folder in each list's directory.

I then invokve MHonArc on the various folders daily under Cron.  I
do this, rather than archiving messages immediately as MHonArc is
fairly expensive and some of my lists are reasonably high traffic
(bursts to 100+ posts a day).

You can find the file relevant to the above, encluding my MHonarc
RCs here:

  ftp://ftp.kanga.nu/pub/Kanga.Nu/WebArchives/

A simpler set of RCs that are not dependant on PHP4 etc, can be
found here:

  ftp://ftp.kanga.nu/pub/Kanga.Nu/WebArchives.old/

> Do you turn off archiving in mailman?  

Yes.

> Do you just run a cron job to update the archive?  

Yes.

> In order to have it group things by months (like mailman does) do
> I have to rotate or move the mailbox each month??

Yup.  I have a cronjob move the "current" symlinks for procmail each
quarter (and for low traffic lists, each year).  The MHonArc scripts
internally calculate the quarter directory names when archiving.  I
then have a little PHP-based index which represents the sections up
in a unified manner ala (shows all the periods archives, how many
messages in ach, links to the author/date/thread indexes etc) as
shown here:

  http://www.kanga.nu/archives/MUD-Dev-L/

and here (lower traffic list):

  http://www.kanga.nu/archives/Noted-L/

-- 
J C Lawrence                                       claw at kanga.nu
---------(*)                          http://www.kanga.nu/~claw/
--=| A man is as sane as he is dangerous to his environment |=--