Hi folks, I've got a head scratcher here and I'd really like your help, please.
The problem arose when we replaced a RedHat 6.2 server with a Mandrake 8.1 server. On the new Mandrake server, a few clients (out of about 400 or so) are now having difficultly POP'ing their mail where they didn't have before. What happens is that they are experiencing an extended connection time, up to 60 seconds before the mail transfer starts after starting a mail run via pop. In some cases, this is causing the client to time-out and report the server is unreachable. If I monitor /var/log/secure and /var/log/syslog while one of these users does a mail run, I see the POP3 connection made almost instantly in /var/log/secure (so it's not a net latency). However, the pop service init login doesn't show up (in /var/log/syslog) until quite a while later; when it does, the user gets his mail. Thus the delay is occurring between the pop connection and the pop login. I was at one client's machine this weekend, trying to figure out what is going on. He has one of those 4 port Linksys home router/NAT firewall units. If I bypass the Linksys router and connect the cable modem directly to his machine, the POP latency disappears and mail transfers are fast. Put the Linksys router in the path and the long latency re-appears; the log entries in /var/log/secure and /var/log/syslog are almost simultaneous. (I do not know whether all the clients having problems have such a router/NAT firewall or not). So far, I've not found out what happens on a POP connection after the initial connection via xinetd and I have no idea what can be causing this delay. Does anyone have anything to suggest? Thanks! -- Cheers, Rob
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
