Revert techniques

Hello everyone,

Sadly in may region there is a recurring vandalism activity. Usually changesets occour in clusters of 5-10, making a response rather urgent before some mapper add further changesets to patch damages. IMHO we should not charge DWG of such local hassle, asking for a revert that, depending on DWG workload, could be performed days later.

So I’m looking for some clarification, since, trying OSM revert scripts [1] “”, execution log reports for instance:

cannot load version 2 of relation xxx (get): 404 Not Found
cannot load version 4 of way yyy (get): 404 Not Found
cannot load version 1 of node zzz (get): 404 Not Found

Until now, revert has been done in “dryrun” mode, so I wonder which results should I expect running effectively.

Should I go for simple reverts, starting from newest changeset? I tried this also, but several conflicts took me time, wondering if it were more practical manual adjustments and more OSM verisions…


Check which API you’re pointing at -it might be the dev API rather than the live one. Testing in non-dryrun mode on the dev API first is actually a good idea - go to , sign up, edit some data normally and revert it.

If not that, check you’re accessing https rather than http endpoints.

I have the same problem.

The dryrun=1 will run and with dryrun=0 I have a problem.

1 object IDs read from complex_revert.log - will not touch these again
cannot load version 1 of node 8189802462 (get): 404 Not Found
cannot comment on changeset: 404 Not Found

I use apiurl=

I would be happy if someone could help me.
Thank you.

surveyor54 is the test server, does node 8189802462 exist there?

I think there is a mistake in the REAME file.

If I use the normal link “apiurl=”, it works again.

Thanks for the explanation.

I suspect it’s deliberate, to suggest that you test on the dev server first.

So this is meant for practice.

OK, I understand.