12.04 server going offline

[SabreWolfy]

I have a headless server running 12.04. On occasions (after several weeks of uptime), I am unable to access the server (via HTTP or SSH) -- the connection times out and cannot route to the server. I hard-reset of the server solves this -- after a hard-reset I able to once again access the machine via HTTP and SSH.

Once I have access to the machine again, what log files should I look at to try to diagnose what is causing the problem?

[chadk5utc]

Hello
I would check logs for errors however Im more inclined to think hardware/wiring issue, I would also start checking the local network. How many computer in the network? how are they connected? Router? Switch? process of elimination check all connections CAT5 to the server/router/switch check the NIC.

Chad

[OliverHaslam]

It'd be interesting to see if the machine has the same IP address after restarting, assuming but I suspect you're not using DHCP anyway.

I have a similar issue where all my machines stop being able to access the web because my DNS serving VM needs rebooting. Are you trying to connect via an IP or a hostname?

[SabreWolfy]

Hello
I would check logs for errors however Im more inclined to think hardware/wiring issue, I would also start checking the local network. How many computer in the network? how are they connected? Router? Switch? process of elimination check all connections CAT5 to the server/router/switch check the NIC.

Chad

There are many other machines on the network, none of which experience this problem AFAIK. The server is less than a year old, but I guess hardware failures can happen at any time. Thanks for the suggestions.

[SabreWolfy]

It'd be interesting to see if the machine has the same IP address after restarting, assuming but I suspect you're not using DHCP anyway.

I have a similar issue where all my machines stop being able to access the web because my DNS serving VM needs rebooting. Are you trying to connect via an IP or a hostname?

Other machines on the network are allocated dynamic IPs, but an IP is reserved for this server, so I've hard-coded it so that it always has the same IP. I tried connecting to the server via the IP and via the hostname, neither of which worked.

The problem is erratic and intermittent, making it difficult to trouble-shoot. I have to wait until it presents again before I can start looking at what the problem might be.

[CharlesA]

I would recommend writing a script that pings, say google or your gateway and logs it to a file with the date/time, so you can track when this occurs.

I would check the sys log for anything related to eth0. I ran into a problem a while ago where my network card on a VM would stop working randomly and I found it was due to a problem with the kernel module for that specific network card (E1000 in my case).