1

I'm using Russell Coker's postal software to stress-test our mail server.

But I can't squeeze more than 24576 messages per minute out of it!

This seems to be a limitation on postal's side as I can fire up another instance on another machine (or the same machine) and easily squeeze through another 24576.

[root@client test]# postal -m 32 -M 4 -t 256 -c 128 -r 0 10.0.0.232 user-list
time,messages,data(K),errors,connections,SSL connections
14:21,13736,254013,0,416,0
14:22,25176,465690,0,358,0
14:23,24576,454808,0,352,0
14:24,24576,451941,0,375,0
14:25,24576,453869,0,376,0
14:26,24576,455206,0,385,0
14:27,24576,452850,0,356,0

Any idea if there's something else I'm missing or if there's something else I can change? After source-diving, I observed the throttling code doesn't even run when the rate-limit is the default (0), so it's not that.

3
  • 24576 = 3*8192 or 24*1024. That seems fairly significant.
    – Hyppy
    May 20, 2011 at 18:47
  • Yeah, it is a rather nice round number.
    – MikeyB
    May 20, 2011 at 19:02
  • Have you tried running another benchmark? e.g. postfix.org/smtp-sink.1.html May 20, 2011 at 23:34

2 Answers 2

1

I ended up writing a python script that would spawn a number of postal processes and aggregate the results.

Sometimes it's not worth solving the actual problem, but putting in a sensible solution that works around it.

If anyone happens to need to solve this particular problem in the future, get in touch with me and I may be able to pass the code on to you.

0

Did you try to give the user running postal more than 1024 file handles?

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .