Opened 5 years ago

Last modified 5 years ago

#5313 new defect

Some buildings / one changeset (12 days old now) is not rendering in osm.org's standard "mapnik" map

Reported by: aseerel4c26 Owned by: Tom Hughes
Priority: minor Milestone:
Component: admin Version:
Keywords: rendering Cc:

Description

e.g. https://www.openstreetmap.org/way/341789642 (and the nearby buildings) is not rendering since 12 days in mapnik. See https://help.openstreetmap.org/questions/42765/a-single-changeset-is-not-rendering-but-subsequent-changes-are

I really cannot see a tagging or geometry problem (closed line, usual tag) and other maps render it.

A /dirty on https://b.tile.openstreetmap.org/19/477089/320402.png does not help.

Is there still some rendering toolchain hiccup due to the server problems on the day of way creation? This way https://www.openstreetmap.org/way/341789642/history even was edited after the problems were resolved (apparently – at least for me updates were as usual at this time).

Is it a single problem, or may this problem also appear at other places of the world / other edits from that time? ... maybe there would be a one shot batch solution.

Change History (3)

comment:1 Changed 5 years ago by scai

Is it a single problem, or may this problem also appear at other places of the world / other edits from that time?

There is an older but similar question at https://help.openstreetmap.org/questions/41966/old-and-deleted-data-gets-rendered explaining rendering problems around https://www.openstreetmap.org/#map=17/50.65175/13.25163. These changes are several weeks older. If you open the data layer or an editor then you will see that the broken landuse rendering doesn't correspond at all to the underlying data. Also, this just happens on the standard layer. Other layers are fine.

I can imagine similar issues all over the world which just doesn't have been spotted or reported yet. Possibly due to various database hiccups and lost diffs?

The interesting question is how we can identify and avoid such issues in the future, if possible at all.

Last edited 5 years ago by scai (previous) (diff)

comment:2 Changed 5 years ago by aseerel4c26

thanks, scai! The "lost diffs" explanation sounds good, partly. But why do subsequent edits to the way not fix it? Because the way is not existing in the database and the diffs to this way are discarded?

So, according to Simon's answer at the other help question "this will be rectified if and when the tile server databases get re-imported". Whenever/in which frequency this will happen. But if it will at some time, that's fine then. Just a really confusing situation in the meantime.

Last edited 5 years ago by aseerel4c26 (previous) (diff)

comment:3 Changed 5 years ago by scai

The "lost diffs" explanation sounds good, partly. But why do subsequent edits to the way not fix it? Because the way is not existing in the database and the diffs to this way are discarded?

Unfortunately I have no idea. I also thought that touching the ways again will fix the problem but it didn't. And I'm lacking a deeper understanding of all the internals here.

So, according to Simon's answer at the other help question "this will be rectified if and when the tile server databases get re-imported". Whenever/in which frequency this will happen.

As far as I know there are no scheduled re-imports. So these problems could persist and accumulate for a long time :\

Note: See TracTickets for help on using tickets.