On 26 Jul 2014, at 01:42, McKinnon Chris <crmckin...@shaw.ca> wrote: > I’ve done some testing with swaks trying to track my performance issue. I > don’t think this is a postfix issue. It is just the most apparent symptom. > I’ve also noticed SSH to my server is quite laggy but if I use the command > line with Screen Sharing it is responsive. Same with SFTP, sometime fast and > sometime really laggy. > > I’ve turned on verbose logging for smtpd and tis but this hasn’t revealed > anything yet.
Because, as you said, this is unlikely to be a Postfix issue. > I’ve been hitting the mail server using swaks remotely vs. directly on the > server, I’ve attached some logs if you are interested. Initially the remote > attempts would take 5-30 seconds from start to finish. Directly on the > server would take tenths of a second. I’ve noticed that the longest waits > appear to be at "220 2.0.0 Ready to start TLS” and "EHLO [10.0.1.7]” but I > couldn’t find any red flags in my logs. Then I found this archived posting > http://www.irbs.net/internet/postfix/0501/1437.html, noticed that I was > blocking ICMP traffic, and enabled it. After enabling ICMP traffic I was > getting external response of 1-3 seconds for a short while but then they shot > back up to 20+ seconds again. Please don't post plain text in formatted documents, it is quite useless. Again, you have shown that this is unlikely to be a Postfix issue. Turn your attention to lower layers, and verify they are working as expected. This means turning off the software firewall you mentioned, looking at the networking itself, possibility of packet loss somewhere, or hardware failure such as faulty cables, a network switch that causing problems, and so on. Or maybe it's the network port in the server itself. Either way, off topic for this list. Good luck! :-) Mvg, Joni
