I've been running CentOS 3.1 successfully. The system failed today. When I rebooted, Linode reported "success", but I could not even ping the IP. I then deployed the small Debian distro successfuly, and used this to see /var/log/messages

The failure appears at these two lines

Sep 5 14:40:45 plato portsentry[1165]: adminalert: ERROR: No UDP ports supplied in config file. Aborting
Sep 5 14:40:45 plato portsentry[1165]: adminalert: ERROR: could not go into PortSentry mode. Shutting down.

If this doesn't reveal anything, can I be certain the problem is in theOS? If so is there an easy way to replace it?

Any help would be appreciated

Have you got your Linode set to use the latest available 2.6 kernel? If so, maybe the kernel version has changed since your last boot and the new one is somehow incompatible with your configuration.

The 2.4 kernel hasn't changed since January but the 2.6 kernel changes quite frequently. The latest one (2.6.12.3-linode14) became available on 20th July.

If you are using 2.6, make sure you haven't inadvertently recreated some or all of /lib/tls.

hth

_________________
/ Peter

I've since discovered the Linode shell. I found that I can log in and run 'top' which shows proccesses.

I still can't 'ping' the system. What should I look for?

A clue might be that the host was recently change to non FQDN.

Many thanks.

Can you ping other machines from inside your Linode?

Since portsentry produced error messages at boot, try disabling it and rebooting. If the problem isn't portsentry, check to see if your eth0 config got broken during the system failure.

_________________
/ Peter