tHog

DIARY 2013

(2012)

DECEMBER

NOVEMBER

OCTOBER

SEPTEMBER

AUGUST

JUNE

MAY

APRIL

MARCH

FEBRUARY

JANUARY

2013

Sat, May 25

<17:33 EEST> Tyng├Ąt finished last night with a blast. We had suffered a few cancellations due to illness and lack of audience, but the final couple of nights were packed, and the show just kept on improving.

Just in time for the final, I think I finally nailed Willow's keyboard issue. The earlier problem had recurred, and I took another look at the keyboard cable along with its potential short circuits. Armed with proper electrical tape, taping all over, it was still not enough. Then I realized there is an unrelated antenna cable dangling in the service hatch, just in the right spot to hit the keyboard signal pins (on the other side of the connector). Having that one insulated, the machine has been rock solid since, and I am pretty sure of it after some mechanical stress tests.

Of course, the RF cable, intended for an extra wireless mini-PCIe card, should already have been protected in the first place. Another cable next to it had proper heat-shrink tubing around the connector, so perhaps this one's prophylactic had simply been lost in some former surgery. In any case, such a dangling bit would explain why the keyboard was often buggy after being suspended for transportation, but would recover when rebooted on a stable surface. The short-circuit would do its damage when moving during transit, with the keyboard controller having power on. On a desktop, though, it would generally hang away from the PCB.

At this point, after over a month of problems, I naturally feel a little stupid having not considered this cable before. There were countless accounts to be found about other possible short circuits with the keyboard cable, and every time I opened the service hatch, I missed the cable while it was there in front of my eyes.

It had also been a frustrating time with such an ephemeral problem, with the keyboard working most of the time, but sometimes going haywire at a critical moment. This had a surprisingly strong psychological effect on me, since the keyboard was an important means of expression for me. I even felt somewhat paranoid at times when the keyboard started typing on its own; on other times it would not register any touch, and I felt equally helpless.

Naturally, I also started to question my usual trust for Thinkpads, the known workhorses of IT. However, the experiences with warranty service and the way I could fix this myself have probably made my faith even stronger; it's not just some shiny piece of iHardware welded shut, but a real machine you can take apart and rebuild as needed.

Sun, May 5

<19:51 EEST> Fritz Post! A rather lame paraphrase of the Slashdot meme for an occasion not too special, but it had to be said.

A couple of weeks ago I had gotten myself a Fritz!Box 7390, as I wanted something more stable than my ISP-provided VDSL2 modem. Gigabit ports were a must, and this happened to be the cheapest one in the range. Unfortunately, getting it into a bridged mode for passing through public IP addresses turned out extremely tricky. I basically gave up for a while, with studies and theatre taking up time and energy. That while gave me an opportunity to step back and wonder if a NATted setup could in fact be better for me.

My networking needs are somewhat complex and conflicting already. This time, the most acute reason for getting another modem was mining Bitcoins and other cryptocurrencies. I was tired of the Elisa modem crapping out on its occasional auto-updates, and I wanted more stable Internet connections for all of my rigs.

So far, most of my machines have used NAT via one of my publicly addressable machines. There are obvious stability issues whenever I have to reboot one of these; I generally have two of them, but it takes some fiddling to update the routes in the local network. Now I simply NAT everything via the modem, and reboot any machine as needed, without affecting the others.

Naturally, this way I lose some of the flexibility of public IPs, but in practice I can still serve everything as before, using port forwarding. In fact, it allows me to distribute services on several machines more easily. Also, with the increased value and impact of cryptocurrencies, I do think about security more than ever. While I can generally manage a publicly addressable machine, a separate firewall with further limitations should not hurt. There is an overall feeling of losing a little control to a black red box, but frankly, I have better things to geek out on than setting my own iptables.

While the theatre piece has settled down to regular performances, it is still taking non-trivial amounts of time and energy. The past week, in particular, turned out surprisingly social, even if I did my best to avoid the usual Wappu celebrations. In fact, I only went to one such event, since I was asked to play a couple of pieces at Viitaniemi school. Of course, it was a nice excuse to see some of my past colleagues and students, and a sober alternative to the overall booze-fest.

I wouldn't mind such social life at this point, but I still have a few important exams coming up. Lectures finished, a lot remains to study and revise. The loss of daily/weekly study structures and the social events have already set me in a kind of summer holiday mood, which I have to fight for a couple more weeks. Fortunately, the few days before the first of these exams are free from theatre.


Risto A. Paju