I am currently dealing with an issue where e-mails won't get delivered because exim has too many connections.
I googled around a bit and found out that maybe I should check how many instances are running, and last time I ran into this issue, I had about 140 instances running. After a restart of the service, it was down to around 10 instances for a while, but is now at over 200.
When I restarted exim, I also changed the config to allow up to 200 connections (was 100 before).
I do not know exim overly well (aka I am totally new), so what I wonder now are the following things : 1.) How many instances of exim are healthy? We have a lot of domains on the server but handle most of the emails on external services. I imagine 200 running instances are a bit too much, but maybe I am wrong?
2.) What is, on average, a good value for smtp_accept_max?
3.) If 200 instances are too many, what can I do to find out what causes it to create this amount of processes? I am not really seeing anything in the logs, but maybe I just don't know what I should look for? And also, how can I fix this if it is an issue?
I read something about changing :blackhole in all files in /etc/valiases to :fail, but when I look at a file of one of the domains that use the server as an MX, the file looks like this :
[email protected]: [email protected]
*: accountuser
Might that config file itself be the issue?
200 connexions may or may but be too much. It all depends on your machine hardware capabilities. 200 sounds like an high but still reasonable value to me. Obviously if your server is a 2ghz CPU with 1gb ram and pottery slow disks that may be a problem. A tool you can use in order to get statistics and maybe point out the origin of the problem is eximstats. This tool will parse log files and present some (not so fancy) graphs
http://linux.die.net/man/8/eximstats