So, I have a SPARC that's been running a somewhat bleeding-edge
Debian distribution. At some point I installed exim on it.
A couple of months ago I upgraded libc6 such that exim stopped
working. No big deal; this box only runs exim to forward mail
to my real machine. So I never bothered to fix it.
Today I upgraded libc and exim so that exim now works again.
I have just discovered that for the past few months of more-or-less
continuous uptime I have been running the exim cron job every
15 minutes. Every 15 minutes, it would try to run exim to empty its
queue. This would fail with a libc-related error message. cron would
mail me a message about this. Which would sit in exim's mail queue,
since exim wasn't working to deliver it.
I currently have more than ten thousand messages in exim's queue.
D'oh!
Peter Maydell
[eximon takes a looong time to start up with that big a queue :->]