RE: Listserv Maximum
Is there a listserv maximum number?
"You grow up the day you have your first *- Terri Slater real laugh - at yourself. *
- Web Programmer Ethel Barrymore *
- Washburn University *
- 785-231-1010 ext 2302 *
ts> Is there a listserv maximum number?
I don't understand the question. Is there a maximum number of what? lists? members per list? postings per list?
There are no artificial limits on any of these. There may be effective limits based on your system and network resources.
-Barry
ts> Is there a listserv maximum number?I don't understand the question. Is there a maximum number of what? lists? members per list? postings per list?
There are no artificial limits on any of these. There may be effective limits based on your system and network resources.
The first limit you hit when sending to a big list with many old addresses (in my case, a list that was never properly cleaned, for lack of software able to clean it nicely as MM does now : ~150K subscribers, ~40K bounces) is the BounceRunner sucking up all your processing power and locking up the database.
You can imagine that bounce processing is something that doesn't really need to happen in *real time*, ie when the server is sollicited by sending the mail. However, if you just collect the bounces and wait before processing them, you end up with high disk space requirements.
So I'd suggest, that we shoud
pre-process bounces without accessing the list db : for each bounce, read the message, extract (when possible) the email address and list name, and write the email address into a listname-bounces.db ; then discard the message
when the mail server seems to get colder (the latest lock on it is older than one or two minutes, or any other activity indicator), load the list (w/o locking) and calculate the bounce scores of the bouncers you have seen, and so on. Then only lock-load it, write the changes, and save.
I'm also thinking:
- should we/how could we do the preprocessing on another machine?
- could Mailman order its send queue by putting the non-bouncers first, thus shortening the delivery time for non-bouncers and enhancing the system response?
-- Fil
participants (3)
-
barry@python.org -
Fil -
terri slater