Strange indeed, it appears the road does not even exist for OSRM. With the mapzen router it works, though. So I guess the problem lies with OSRM and not with the OSM data.
LePe, it works better now than yesterday. I do not know what the original problem was but my guess is that when you or someone else was editing in that area you created and then deleted or moved some additional highway that OSRM is trying to base its routing on. The routers’ databases lag some hours/days behind actual OSM data. The problem should be completely gone in a couple of days.
I found this problem about a year ago. So far I’ve been only waiting for corrective actions. Because I use this route it’s somehow problem to me. Now I’m trying to fix the problem.
This problem exists when using on-line navigation or OSMANDR off-line mobile map as well. They are two totally different navigators and they make exactly the same mistake. So far I don’t know any map apps where this really works. That’s why I think this is something to do with map data and not with map application. I’m afraid this problem is hiding inside the OSM-database.
Yes I tried to fix this problem. To test - I draw an other highway line along with the original line from Lahdentie to Porvoonväylä. it did not work either and I really wonder why.
If you find a solution to this, please help. It would help to hundreds of motorists navigating east of Helsinki.
There have been lots of changes in that area in the past few days. including ones from you suggesting that you have been experimenting on the live data.
I found this barrier on a node of the motorway (with only *=no attributes), where a trail crossed the motorway. I have put a gap into the trail and moved the barrier away from the motorway node.
A barrier on a node can be a reason to remove any way from the internal routing database, which cannot be reached behind the barier or which is a dead end oneway because of the barrier.
It looks to me as though whoever set the barriers is trying to say that the slip-road itself is the barrier, when the real barrier will be some feature on the edge of the motorway, and there may not even be a physical barrier, so the trail should just end on an anonymous node. I think the barriers need fixme’s at the very least.
The barriers are actually the two crash barriers that are either side of the motorway. There are no gaps in the barriers so the trail stops without joining the motorway. Don’t use this information to change OSM database though.