| |||||||||||||||||||||||||||||||
|
(mostly to appease the Yahoo!s). Before I do that, I want to see what the impact is on my mail outgoing throughput. What I want to know is how much time a message of a given size spends going through the milter on average. I can measure it indirectly or directly. Since I don't see any options to dkim-milter's configuration to emit timing or detailed logging information to get this information directly, and I can't see how to make postfix emit the information on how much time the message spent in the milter, I figure I need to do some indirect timing by comparing time to go through postfix with and without the milter. I set up postfix to relay all mail to the smtp-sink, and set up smtp- source to generate messages that will be signed by the milter. However, timing this is tricky... I can time easily how long it takes to generate and inject the messages into postfix by smtp-source, but tracking when the last message leaves postfix is a bit trickier. I can see when smtp-sink has received all of its messages, and I can see in the log file when the last one was delivered to the sink, but then timing becomes a manual process, and the accuracy is limited to 1 second (which I guess I can live with.) If I could make smtp-sink exit after N messages, I could use that event to mark the completion of each test run and it could then be automated. Any ideas on how to automate detecting when all test messages have gone completely through the system? Basically I want to know how long it took from the time smtp-source starts to the time smtp-sink has received all the messages generated. Also, is the milter single-threaded or does it handle multiple simultaneous connections from postfix? I'm pretty sure it is handling multiple messages at the same time but I just wanted to confirm. The docs for dkim-milter are pretty thin. Thanks!
| ||||||||||||||||||||||||||||||
© 2004-2008 readlist.com