Frequent offline debian apache2 server
 

Hi All,

I have a debian (wheezy) web server (home built machine) that frequently (once a day) goes offline mysteriously. I administer the server via ssh and when it goes offline a "Broken pipe"g message appears, which means that the SSH connection died. Websites are no longer reachable. If I connect a monitor to the webserver after the server is unreachable by SSH and sites unreachable, the monitor screen remains black (entire system is dead?). The only thing I can do is a force reboot via the power button. Many times it takes two reboots for the server to become reachable again.

I have reviewed /var/log files and searched and have come up empty on a possible the cause.

When this happens, I notice that /var/log/syslog file reports one or more tables in a mysql database has crashed and needs repair. However, repairing the database tables via myphpadmin does not prevent the server from going offline again, which may indicate the database table crash was due to the force reboots.

I have checked the server for rootkts with rkhunter (1.4.0) and none are found.

I can post only 2 minutes of the syslog file and doubtful it will be very informative.

Any suggestion on what to look for in my log files or any kind of pointer will be appreciated.

System: Debian (Wheezy)
Services: apache2, fail2ban, squirrelmail, pure-ftp, courier, bind, autofs
Hardware: 1.7 Ghz CPU, 2Gb RAM, 1TB disk space,

Update (Oct 14): Apparent cause was postfix not running, see post below dated Oct 14. Started the service with the /etc/init.d/postfix start command.

is there any message at the first attempt to boot?

With the information available it sounds like a hardware issue. Heat?? Ram?? Had that happen with an old IBM 300 and it turned out to be bad ram.

Thank you tux11 and leandean for the replies.

I noticed that postfix was not running, so I started the postfix service. Since starting postfix, the problem appears to have disappeared. I say disappeared because I feel like I have not done enough to know the absence of postfix was causing the problem. The system has been running for the past 30 or so hours, which is significantly longer than the 1 hr or less that was happening.

For now, I will call this problem apparently solved.

Thanks again.

[SOLVED] Heat was the cause
 

This is just a follow up post with the final solution that was causing the problem described in the first post of this thread. The root cause was heat, which was due to a faulty power supply. The symptoms described above has gone away after replacing the power supply. The system has been running for over a week without issue.

Thanks tux111 and leandean for replying. The replies helped pave the way to a resolution.