[Mailman-Users] Automating Mailman Archive Maintenance

[Mark Sapiro]

C Nulk wrote:
>
>I am in the process of trying to automate our Mailman Archive
>maintenance before it gets unruly.  I looked in the FAQ and wiki for
>information and found some about rebuilding the archives (which will be
>handy) but nothing about automating it.
>
>The assumptions I am working under:
>1. the html files for the archives are located in <some
>prefix>/<listname>  (to be called DIR-A)
>2. the directory the mbox file to rebuild the archive html files are in
><some prefix>/<listname>.mbox  (to be called DIR-B)
>3. our automated process will process the mbox files in DIR-B and delete
>completely or mark for deletion any messages older than a given timeframe.
>
>Now, the questions:
>1. If I run bin/arch --wipe <listname> to rebuild the archives for
><listname>, do I have to delete the files in DIR-A first or will
>bin/arch do it?


You do not have to delete any DIR-A files. That's what the --wipe
option does.


>2. When a message is added to the mbox file in DIR-B, is it appended to
>the file or does it get added through some interface?


It is appended.


>3. When a message is added to the mbox file in DIR-B, are any existing
>messages that are marked for deletion removed or is the message just
>added to the mbox file?


It is just appended by a file open and append operation. The process
does not in any way emulate an MDA or any IMAP or other mail access
type process.


>4. When bin/arch is run and builds the html files, does it ignore
>messages marked for deletion or does it add the message to the html
>files no matter how it is marked?


It totally ignores any message status type headers.


>5. Should Mailman be shutdown prior to running my automated process,
>which includes running bin/arch, or can I leave Mailman running?


It's OK for Mailman to be running. There are archive locks that will
prevent concurrent updates.


>6. In our installation, the public archives directory for each list is a
>link to the private archives directory for each list, is that the
>standard or should I be prepared to see some archives in the public area
>and other in the private depending on the particular list's setting?


All archive data is in archives/private/. archives/public/ contains
only symlinks.


>7. Is there any other gotcha I should watch out for when using an
>automated process?


I don't think so.

-- 
Mark Sapiro <mark at msapiro.net>        The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan