The weak link
So, it turns out that my gateway computer (which is more than just a gateway, but it doesn't matter) died this weekend, making the whole of my LAN unable to get online. The culprit turned out to be the motherboard, that had been showing signs of old age for a while and probably the power supply too, since it had started to make a bit of noise. Also, it looked like for a while like the hard disk was dead but it turned out that it was still going. (Even so, it was replaced, since it was getting noisy.)
So, I had no other choice than ordering new parts and wait for the few days needed for them to arrive. The majority of my LAN stayed disconnected from internet during that time, mostly since setting up another computer as gateway/firewall temporarily would have been too much work. So only our laptop, which does have a decent firewall, since it's sometimes on untrusted network, was allowed to be online at that time.
Now, I got the new parts Thursday and got the whole thing running again, after transferring all the data from the previous hard disk to a more reliable one and recompiling the kernel to accomodate the new built-in network card.
The lesson to learn here is pretty obvious. Having a single point of failure like this, that can take down the whole LAN with it, is a pretty bad thing and in a professional environment, there would probably be at least one computer ready to take over if one fails. In practice however, I don't have much space or time required for one more computer, so the best I can do is to have a backup solution for the most important network services when something fails on my side. Emails are already taken care of by a backup MX (Which is ran by a friend of mine), but at least the webserver would benefit from having such a fallback as well.
These are plans for the future, anyway. The positive part of that outage is that I was able to work on other projects for a few days
Comments