Bit of a gap since day 18 I know, but work and organising openstreetmap stuff got in the way of working on openstreetmap. So, today lets try to get some rails done in the calm between bugfixes;

• Moved the planet generation back to 1am rather than 4am which was for db problems last night
• uploading new planet in to the (mapnik) tile server. For some reason it’s not putting rendered tiles in the db since the engine change from myisam ti innodb, investigating. Hrm the filesystem has millions of directories for the temporary files… deleting…. new planet now imported… tiles seem to be rendering but there’s a backlog.
• Fixed the mime type bug on planet.openstreetmap.org, now figuring out why I can’t log in to trac. Ok fixed the login bug. closed bug 383 (mime type)
• getting stevecam back up, waiting for tile server to load up new planet. Quick recompile, stevecam back up!
• So it turns out the tile weirdness is because people have been trying to spider all the tiles. But, since many tiles don’t exist they get entered in the dirty queue to render.. which just expands… and expands. I’ve blocked the ips responsible and am now implementing a limit. 🙁
• lunch
• Limit implemented. It’s now 2,500 tiles / day / ip. Also, tiles which havn’t been viewed in more than a month (28 days) are deleted every day. See stuff over here for implementation. Cool all seems to be working.
• deployed new openlayers, closes #310
• replying to a ton of mail, organising various conferences and stuff I’m speaking at for OSM…. Added to wiki events page
• Some final food for thought. Here’s a histogram with number of ip addresses on the y axis, number of tiles downloaded for the past 5 or 10 hours. As you can see the vast majority of people download under 2,000 tiles. But we get the odd person grabbing 10,000 or so. What should the limit be?

• Load up the new planet… done. New tiles rendered should use full world vmap0 layer and correct pri/sec/trunk road colouring
• And again today – london party stuff is showing up
• conf/party organsising stuff and then IRC SOTM meeting – going to be tons to do
• put the hacky tile autogen stuff in svn look here but its very very hacky

• Synced the tile generation xml with upstream (fixes a/b/trunk road colouring)
• installed apache2-mpm-prefork on tile. -worker causes problems with multiprocessor machines and mod_ruby (fixes broken tiles not downloading)
• removed the old mapnik layer and wms-c. Moved mapnik tiles to a nicer place (not /test/).
• some mapping party organising (see following post in just a sec)
• Brought the rails front page up to what the current one looks like.
• Replied to a ton of email and did some conference organising stuff – looking good
• Initial steps of importing an OSM dump in to the rails port to see what’s left to do and if it succeeds.

• Brought the mapnik db map layer on the front page up. It automagically renders data if it’s not there or if it’s too old.
• done some more organising of the various upcoming parties, see previous post for one bit
• deploying Nick Burch’s applet fixes… done. Also added a little ant task to sign the jars.
• edited down the geospatial vision podcast (see next post). Might be able to get some audio out of talking to people at the pub at the oxford meetup too, not sure yet.
• good IRC meet about the OSM international conference organised by andy robinson, lots of little jobs to fulfill!

• Done lots of organising for the 3 upcoming parties and the conf over the past week.
• Figured out why the gpx import keeps stalling – someone keeps uploading mpegs which the importer thought were xml files in a strange charset, then crashing when it couldn’t parse it – fixed
• Brought back the wiki machine… seems the VM just isn’t up to it any more, will need to think how to proceed

• Spent the morning trying to build ruby bindings for mapnik, didn’t get anywhere, mailed mapnik-devel. update short debate on mailing list, looks easy if you know C++, mapnik, swig and python well enough but I don’t. Someone please make swig bindings for this and I’ll have your babies.
• brought back stevecam. It was pointing out my window and I forgot about it sorry

What a morning! Spent it all organising various parties and the up coming conference. There’s now a party mailing list – if you want to help organise or run a mapping party please subscribe there. It’s a transitory list so feel free to just subscribe for the period of organising the party you are interested in! I’ve deprecated the old talk-london/midlands and iow lists. I’ve done some tile stuff too – they’re rendering now whilst I do other things.

Lucky 13 begins

• Told list maintainers that they can change the language of their list pages (result of yesterdays stuff)
• Doing tile stuff (they’re getting out of date, time to fix them)
• Various party organising things (we have a room for sheffield thanks to sheflug it looks like!)
• Went out and did some mapping whilst waiting for some tiles to generate (see below). Accidentally deleted the trace off the GPS. d’oh!
• Tile stuff done for today. There’s a new layer on the map, ‘Mapnik db’, which is todays planet dump. All the tiles come from a database and 3 CPUs are rendering every tile from zoom 0-18 as I type. Currently it’s good down to about zoom 11, we’ll see where it is tomorrow.

Done!

Let’s-a-go

• restarted gpx import
• submitted an xtech proposal. And a mapping party tutorial proposal.
• copying over a ton of tiles that RalfZ has rendered to the main map (in progress)
• rm’d all the blank tiles with find -size 158c -exec rm ‘{}’ ;
• Many updates on the applet due to Nick Burch, try it out!

done!