Hi!
At some point in time yesterday one of my test servers (let's call it
"ringo") got overwhelmed with emails and ran out of disk space in /var
partition. I had more than 75,000 emails at some point with 95% of them
not pre-processed.
After that, some weird issue with the way qmail-smtpd logs messages
showed up. I started getting entries like these in my "current" log (I
run qmail with multilog through tcpserver):
qmail-smtpd 3568: message queued: 1125024832 qp 16082 size 118074401 bytes
The weird part here is the size of the message. 118074401 bytes is off
the charts. databytes is set to 18,000,000 bytes. And the actual
messages were 5KB to 100KB in size. So, the size of the message as
reported by qmail-smtpd looks really strange. All the servers that were
sending email through "ringo" were running sendmail version 8.13.4 or
8.12.10.
So, the question is - is that some kind of hard-to-reproduce bug with
qmail-smtpd that shows up only when it runs out of disk space or does it
have something to do with senders (sendmail)? At the end of the day all
the messages were delivered correctly, so I don't understand why logging
was incorrect.
Also, unfortunately the logs rotated so fast when disk space became
available again that I can't track it back to the moment when this issue
just appeared for the first time.
Any ideas?
Thanks in advance.
