On 10/27/05, Bob Dodds <[EMAIL PROTECTED]> wrote:
> TipJar::MTA appears to run on one machine and replace
> qmail. It handles temporary delivery failures by exponential
> retry delay. I'm not sure I need to replace qmail-inject and
> qmail-queue for a non clustering situation. Maybe you'd
> like an all-perl system? I haven't noticed any problem with
> qmail's retry schedule. As far as breaking out
> TipJar::MTA::queue, we don't need to interact about
> headers, we're done, queue it. I don't think so.

just qmail-send.  For a clustering situation, TJMTA just
might work fine as is, with multiple outbound boxes sharing
a network drive.  The current architecture of forking
multiple processes which each work on the next sendable message
that no other is working on should map to shared drives with no
modification, as all IPC between the sending processes is through
the filesystem. Give your TipJar::MTA instances on each machine
the same MTA root directory and I bet it will work, until the network
shared drive becomes a choke point.  How much outbound volume
are you dealing with that you are looking for clustered outbound MTA
rather than simple round-robin?

The future direction for TJMTA is using a directory rather than one file
for each message, in order to have multiple recipients, keep statistics,
etcetera without needing to touch the message file.  If I ever get around
to doing that.

Reply via email to