I've configured in our exim configure:
::::::::::::::::::::::::::::::::::::::
remote_max_parallel = 5
...
smtp_accept_max = 100
smtp_accept_queue = 20
smtp_accept_queue_per_connection = 20
30min ago I had checked our test mailrelay
exim + amavis (perl-11) + sophos
which has a amavis-transport to check every mail for one
of our 4 MTA in domain urz.uni-magdeburg.de.
Therefore I stopped our fast primary mailrelay!
Yesterday I had realized the same attempt, but with other
accept-options:
smtp_accept_queue = 99
smtp_accept_queue_per_connection = 99
It was a "good" crash....after 5min I had more than 90 amavis-processes
and ... also sophos-clients (called from amavis) and more than 100 exim-
processes.... and that at a host with only 256MB Memory...
I'cant do anything ....
and today with the reduced options:
smtp_accept_queue = 20
smtp_accept_queue_per_connection = 20
After 5min I had more than 60 amavis-calls ...(and it goes up...)
(amavis will be called only in 1 transport which delivers email to
our 4 urz-MTA)
I had thought, that with the option
smtp_accept_queue = 20
only (max.) 20 delivery processes can work parallel ...20 x amavis...
Where is here the mistake ???
MfG Margrit Lottmann
+-----------------------------------------------------------------------+
| phone : +49 391 67 18572 |
| +49 391 67 12819 |
| Margrit Lottmann fax : +49 391 67 11134 |
| |
| Otto-von-Guericke-Universitaet Magdeburg |
| Universitaetsrechenzentrum, URZ-N |
| Universitaetsplatz 2 |
| D-39106 Magdeburg, Germany |
| |
| Margrit.Lottmann@??? |
+-----------------------------------------------------------------------+
