I'm not sure what the best forum for this is, so Mr. Moderator, please feel free to move to a more appropriate place.

Over the weekend, for the 2nd time, my system "kind of" locked up. Both times this has happened, it's been after around a month of up-time, and both times were at odd hours, so the system wasn't being used much, so I don't think it's temperature related, otherwise I'd get them during the day, when I am using the system heavily.

OK, the symptoms are this. On the monitor I can see my screensaver, but it's not "in motion". The keyboard and mouse don't work. I can't connect by ssh or VNC, but neither actually time out, they just hang. Requests to DHCP hang. BUT, the system is still running, because it acts as a server/gateway for the rest of the machines in the house, and they can connect to the internet without problems. I can surf the web, I can send and receive e-mails, anything but connect directly to the server.

The only way to get the system to respond, is the "big red button". On restart, the only strange thing I noticed, both times, is hundreds and hundreds, if not thousands of messages showing Reiser replaying transactions. It takes nearly 5 minutes for this to complete.

Examining the logs doesn't show anything, except that all logging appears to stop at the same point in time, which I think is when the "issue", whatever it is, occurs, because the time was around 4 hours before I noticed the lock-up.

Any thoughts on what might be causing this, and more importantly, how to track down the culprit.

Cheers.

It is possible that the hard drive has some bad sectors.

When this happens, any I/O is frozen and most applications will stop working.

The kernel keeps running, which is why network communications can still pass through the box.

Services such as SSH fail because they do have to do a little I/O, like read /etc/passwd.

This doesn't answer why the keyboard and mouse stop working, though.

Is there any possibility, whatsoever, that this box is locked hard and your web/e-mail is going through a wireless router or somesuch? That would make so much more sense.

The only time I have ever had a system lock up as you describe it, is when the hard drive starts to show signs of failing. I've never seen the ReiserFS wailing though, when something like this happens (I've only switch to ReiserFS as of a week ago).

However, when the failing hard drive locking my system up became more and more frequent, I would boot up, and immediately pop over into TTY1. When the hard drive finally failed, the ext3 handlers would print messages concerning not being able to access (read/write) inodes, and then the system would grind to a dead halt.

Nope. It's a wired system.

I thought about a drive issue, but smartd isn't reporting any problems, normally, or via short/long self-tests. Plus, I'd expect some "failure" messages at some point, not just a lock-up.

Cheers.

I wouldn't know the first thing about ReiserFS or any of that stuff, but I may (possibly) be able to explain the network anomalies.

Depending on your setup, I'm pretty sure that a "middle-man" computer that connects another computer to a network doesn't even have to be switched on to carry the signals. I know that, back when we had dial-up, I used a modem (inside a computer with no power supply) to extend my phone lines :> Not sure if this would hold true for Ethernet, though.

However, (again, depending on your setup) an Ethernet card has the capabilities in its ROM to forward packets not meant for its workstation. So, even if your machine froze, the card could still forward packets back and forth around the network.

Nope. Doesn't work like that. The server has to be on, and running, to move the packets received on one NIC to the other.

Cheers.