Archive for September 30th, 2008
dovecot problem
At least, that’s what i thought first.
I have this server on my hands, i installed apache, php, mysql, postfix, dovecot, amavis, spamassasin and clamav. Worked great with the versions i installed first, but now has some problems with dovecot server – pop3/imap. I search the logs, even put a post in dovecot mailing lists. And since my first post, and my first reply, i started looking deeper into this shit.
I discovered that on sept 28th at 2:30am, dovecot login processes died suddenly. I also saw in mysqld.log that mysqld died around that time. So i thought, that dovecot being tied to mysql, that should be the problem. So, i manually stopped and started mysqld after 5 minutes – dovecot login processes were up and running. I looked into messages log, ..nothing. I looked into maillog, and saw that at 2:30am on sept 28th, postfix died and was restarted few minutes after…and so was amavis and spamd. Curious…who the fuck sends sigterms to my processes?..So i looked more in logs. I saw in secure log that even sshd was restarted at about that hour. So, my guess was that the server was restarted, or at least, because the boot log was empty, something was done to the virtual machine on which it’s installed. So i typed “w” on that server and saw that uptime was like an hour after my logs said that all those processes were stopped. So, now, my guess is, the guys hosting the virtual platform, did something to this server. I’m gonna report this to the owners and tell them what they should tell the providers. Hopefully, they’ll omit words like shit, fuck, crap… 
