My Site Was Down For 4 Days.

I went away on a camping trip, from Monday, July 4, through to Thursday, July 7, inclusively. I had left some of my computers running on auto-pilot, hoping that they would continue to do so for the duration of my absence from home.

However, close to 8h00 on July 4, which actually means on the very Monday Morning I left home, there seems to have been a power failure at my home, which took down this server (temporarily). This means that my site was actually gone for the entire duration of my absence from home.

Even though no corruption or loss of data took place, I needed to be back home and in front of my computer, before I could restore this site. Therefore, as of 18h35 on July 7 (Thursday), the site is back up.

I actually heard that the power outage in question was an unusual one, even for Hydro-Quebec, as evidenced by the fact that this time, the power was reported as out for a duration of 3 hours, by certain people who I trust. Usually, the way outages due to thunderstorms work, a circuit-breaker in the distribution lines tries to make more than one automated attempt to turn the power back on. This relies on technology which has existed for decades by now, in hopes that a ?tree branch? or a direct lightning strike on the lines, has cleared itself up, and such automated resets actually save utility crews a lot of work and time, who would otherwise need to go out into the suburbs, just to do a manual reset. And then, the first automated reset can even take place within 15 minutes of a minor failure.

Not that that would have restored my site. None of my computers are set up, to a do a reboot automatically, if the power has failed and comes back on. Even then, the computer would still have had to wait for me to come home – 3 days later – to reboot it.

And, since according to one report this outage lasted for 3 hours, it was no routine deal for Hydro-Quebec.

Also, now that I am back home, I see that there has been an update to the “Dyn Updater” Software, which “DynDNS” clients like me can use, so that the IP address which our host-names resolve to, track future changes to the IP address, which my ISP may hand me, as a regular customer.

This update could lead to some better performance in the future, as v5.2.x had failed to track certain normal IP-address changes in some cases, in the past. I now have Dyn Updater for Linux, v5.3.2 installed, and it works so far.

Camping for 4 days was much fun.




I go the unusual step, of hosting my Web site on my private computer at home. This means that such factors as network interruptions and power failures can affect my site, while if it was hosted on a paid-for service, it would probably not be affected by such things.

This morning, at 5h30, I had a very brief power failure at home. It was so brief – less than a second long – that the power supply capacitor in most of my appliances, and in most of my computers, held enough charge to keep those running. ( :1 )

However, the supply capacitor in the computer I name ‘Phoenix’ did not hold enough charge for this, so that ‘Phoenix’ went down. That this the computer which actually acts as my Web server as well.

Therefore, my site and this blog were unavailable until about 5h50 this morning. I apologize for any inconvenience.



1: ) In fact, the compressor of the air conditioner in my bedroom was running. And when the power came back on, its motor had not stopped rotating completely. And for that reason, it was able to kick right back up to full speed.

Normally, if the compressor has stopped spinning completely, an immediate attempt to power it back on will result in a stall, and then in the thermal-trip of the compressor motor, which always has a thermal protection switch built-in.


Why I Am Happy, that my Computers Are Working Again

The recent power failures left me in quite a state of distress, not knowing what the fate of my computers would be.

The computer acting as my Web server, ‘Phoenix’ is a Linux computer, running a “Debian” version of Linux, and a flavor of that, which is “Kanotix / Spitfire” . After the second power interruption this morning, ‘Phoenix’ was actually easier to restart fully, than my Windows 7 machine ‘Mithral’ was. This was somewhat reassuring, since ‘Mithral’ has stronger hardware, and since If the software on ‘Mithral’ was ever permanently messed up, I could in fact try to resurrect it by installing Linux on it. It seems that Linux was after all more stable than Windows.

But what happened to ‘Phoenix’ was also better than a scenario would have been, which I had running through my head between 7h30 this morning and 12h00, the time at which I got ‘Phoenix’ running again.

I had had the scenario in my head, that ‘Phoenix’ could have started to perform an ‘unattended upgrade’, at the moment the power went off, a coincidence which I would have been unaware of.

Luckily, this was not what happened.

But had this happened, my own version of what would have gone wrong differs slightly from the official version, according to which the package manager would simply have gotten jammed in some locked state.

There happen to be other power-users, who complain on the Kanotix user forum, that they had been running a lengthy upgrade while their power was strangely cut. Those people ask for Expert support in unjamming their package manager, which more detached people on the forum give advice on how to do.

According to me, they had such a hard time unjamming their package manager, because this is not all that was wrong with their computers. According to me, those users suffered from two problems at once: A jammed package manager, plus A corrupted file system.

I had a vision of having to approach the Kanotix user forum with the familiar line, ‘An upgrade was running, when the plug was pulled.’ But luckily, no upgrade was running at that time…

…And, there is a specific reason why No Unattended Upgrades Were Running. After I rebooted ‘Phoenix’ successfully, I performed the upgrades manually, which were to have run, just to confirm that my package manager still works 100% .

Link To Previous Posting