At 1:43 PM -0400 6/21/02, Steve Thomas is rumored to have typed:

I just moved mail delivery for my domain from from server A to server B

I suppose this is a really _lousy_ time to note that before you moved a _production_ list, you should have debugged the install on server B with a tyest list or two, eh?

Un/Subscriptions (with confirmations) and digests work, but when a message is sent to the list itself it just goes *poof* and disappears.

Of course, it really doesn't "dissapear;" _something_ happens to it.

LOGABSTRACT is set to yes, but logging stops after the message is added to the digest folder.

Huh? Digest folder, Gracie? It weould probably be helpful if you mentioned exactly what you're talking about here...is this a digested list, or interactive? By "folder," I would assume you mean subdirectory, but what is a "digest folder?" What do you mean, "added" - is the message "added" to archive/latest/digest.header|body, or "copied" whole to archive/volume2002, or archive/latest/, or...or...

Could the problem be with using a newer version of the procmail binary?

The problem could be _anything,_ at least from where I'm sitting, since you haven't exactly given us anything we might be able to work with. Try starting at the beginning, and explaining exactly what course a test posting takes through the install. If you have a digest list connected to it, remove it temporarily. Also, replace your dist list with a "short list" of a couple of your own addresses...or better yet, create a test list and debug THAT. Charlie (who really _does_ want to help, but is frustrated by the lack of information you've supplied)

At 1:43 PM -0400 6/21/02, Steve Thomas is rumored to have typed:

LOGABSTRACT is set to yes, but logging stops after the message is added to the digest folder.

Hum...this can't _possibly_ be right, but would explain what I perceive to be what you're saying. In my connections between interactive lists and digests, I don't follow the "acceptable" method of subscribing the digest list to the interactive one. Instead, I manually pipe the message to `flist list-d` in the interactive list's rc.local.s20...this solves a whole boatloads of problems, not the least of which is piping _before_ list-specific header/footers are added, headers are removed (or lord help us, munged), etc., etc. So the recipe would look something like: # rc.local.s20 # # First, send this to the digest (which will send it back...) # :0c | flist list-d Now, if one were to forget to pipe a COPY, as in: # # First, send this to the digest (which will send it back...) # :0 | flist list-d ...that would deliver the ORIGINAL message off to the digested list before SmartList got around to distributing it to the dist list, instead of delivering a COPY of the message. That would, "stop logging after the message is added to the digest folder" and show the message in the digested list without it appearing in the interactive one (the digest would send it back, but by rc.local.s20 the Message-ID's already been added, so the return would be silently discarded). Charlie (who knows this isn't "the answer," but it bothered the heck out of him anyway)