Hi
Lats week we had minor unexpected power disruption. This cause the system
to reboot unexpectedly. Since then, the mail delivery agent is not
functioning properly. We have SGI Chalage L with dual processors running
IRIX 5.3 with XLV file system. The procmail we are running is 3.10. It has
been running smoothly since we installed it last year.
When the machine comes up again, everything was as it was (we think) no
configuration files was touched or changed and basically we restatrted the
machine so it can cleanly fix any itself (only the mirror mail volume was
offline and we had to put it online manually)
Now, what is happening is really wierd. The procmail can't deliver the
messages and in the mailq we have is huge and its entries have somthing
like this:
--Q-ID-- --Size-- -----Q-Time-----------------Sender/Recipient----------------
DAA03829* 0 Wed Jul 24 03:30
<owner-wwii-l(_at_)UBVM(_dot_)CC(_dot_)BUFFALO(_dot_)EDU>
(open timeout on /home/poststudent)
<pweaver(_at_)echidna(_dot_)stu(_dot_)cowan(_dot_)edu(_dot_)au>
Also the number of procmail processes reached more than 550 process.
At the moment any one try to send message to our system will get this
error in the bounced message:
----- Transcript of session follows -----
UX:sh (procmail): ERROR: fork() failed: too many processes
554 <ekawas(_at_)echidna(_dot_)stu(_dot_)cowan(_dot_)edu(_dot_)au>... unknown
mailer error 1
Another thing, if we send TERMINATE signal to these procmail processes,
the mailq will start to decline until it reach somthing like 40 messages
but then it will start accumilating the mail again!
any clues/help would be much appeciated.
Regards
----------------------------------------------------------------------------
Emad Kawas
Systems Administrator, The Virtual Campus
Email: e(_dot_)kawas(_at_)cowan(_dot_)edu(_dot_)au
Pmail: Edith Cowan University
2 Bradfortd St, Mt Lawley, WA, 6050
Australia
----------------------------------------------------------------------------