Database upgrades

Today I upgraded the slave databases to a new major release of the database engine. I'd like to use some of the new features to increase the performance of the system.

The upgrade itself, and the necessary conversions, wouldn't have caused any outages, since I can tell my software to do queries on another database server while upgrading one (in fact, they will automatically fail over to another server if one is down). But, as usual, there was something I overlooked, and for a few minutes about 50% of the /info/ page requests failed and complained about problems with looking up nearby cities. I had to improve the stored procedures a bit to get them working on the new database engine version.

Please, drop a note in the blog comments, if you notice any other problems.

Account system upgrade completed

The planned upgrade has been now been done. It only took some 15-20 minutes to resolve the few issues that popped up. More details about the changes can be found in the previous blog post.

If you already have an account on aprs.fi for posting AIS data, and you don't have your login password (it's different from the AIS feeding password, and no, you didn't have one before since there was no such thing before today), try logging in with your email address and any bad password, and you'll get to the "forgot my password" path which lets you reset the password to a new one.

As usual, there are quite a few new strings to be translated. From now on, you'll have to sign up for an account and log in to access the translation tool. I hope you don't mind.

Planned upgrade: new account system

It's been a bit quiet here for some time, since I've been working on some larger changes in the aprs.fi back-end system (and learning how to fly, crash and repair an RC airplane). I'm planning to install the changes on the production servers early on Tuesday or Wednesday morning (7th or 8th of July 2009, around 4:00 UTC). The outage shouldn't be a long one, if all goes well.

One of the visible changes is a new authentication and authorization layer, which is based on a conventional email address + password model with a team-based authorization system. It'll enable some new features and services later (once I get around to implementing those) - in the first phase it just lets you sign up for submitting AIS data without manual work on my part.

There's no need to worry, as a normal user viewing data on aprs.fi, you won't need to sign up for an account. Logging in is completely optional, and I won't encourage or beg everyone and their cat to sign up. In fact, the "login with your nickname to see the map" model will go away - you'll end up directly on the map page when you arrive on the site, and the "most usual" site entrance path will be easier and quicker than before.

I will be adding an "introductory" entrance page for first-time visitors later, so that there's a chance to let them know what this site is all about. But it'll only be shown once for each visitor.

New simultaneous viewers record and related slowness

Seems like we hit a new high of over 1000 simultaneous map viewers today, mostly thanks to Dayton Hamfest, and a popular live Hamfest video feed with an embedded map. They're giving away freebies.

A couple of components started hitting file descriptor limits, which had last been upgraded over a year ago. Too many simultaneous connections per process. This made the site perform very, very slowly. I quadrupled the limits, and the site started to perform quickly again, I hope that's enough for more than a year to come. Well, I have to admit that it would actually be a nice surprise if the site would be so popular that it wouldn't be enough...

I also fixed a bug in the "first heard" algorithm pointed out by Ian, VK1IAN. The digipeater alias GATE was not treated as a special digipeater alias (like WIDE, RELAY and TRACE are), and an igate which first heard a packet with a GATE in the digi path was not given credit for hearing it first.

Another fix that went in was a filter which takes out complete APRS packets which have somehow made their way to the comment field of another APRS packet. Apparently something is loosing CR LF sequences between packets (could be my code...), which causes packets to go into the comment of the previous one. Before I find the actual bug I've added a filter to strip these off.

Dayton Hamvention APRS view

Dayton Hamvention is here again on next weekend. I've set up a separate page showing live APRS activity at Hamvention.

Bob Bruninga is expecting to have APRStt running there, too, so if you're visiting Hamvention, you might be able to place yourself on the map using just a DTMF-capable 2m transceiver.

I've also added a new parameter 'he_maptype' in the APRS map embedding interface which allows you to select satellite, physical or hybrid map views in addition to the normal street map.

Date range with weather and telemetry lookup

I'm on a 2-week sick leave (my gallbladder was removed last week), and have some spare time for enhancements and bug fixes. Rest of the time I'm entertaining Armi the cat, who joined us on Friday.

This morning I added arbitrary date range lookups to the weather and telemetry pages. It allows you to look up a detailed graph between any two dates or times. It also has a little calendar widget for your convenience.

I've also done some XHTML validity fixes for better browser compatibility, switched to a smaller position marker on the static maps, and added some navigation links in the bottom of most pages. I also fixed the bug of date selection menu not appearing when clicking on 'start tracking' on the live map.

Also, the /info/ page of a weather station now shows a brief 'latest weather' report instead of the number of weather reports in database. It's probably much more interesting, and it's much quicker to load, too, since it's almost always cached in memory and doesn't require a potentially large disk read.

Google Maps loading problem fixed

Google released Google Maps API version 2.151 today, which somehow broke map loading for many users (including myself). I received a number of complaints that the map doesn't load. The Firefox error console gave the error "window.jstiming is undefined".

I've now manually downgraded to Google Maps API version 2.150 which seems to work better.

Sorry for the trouble, and thank you for your patience!