This street was also not mentioned in my notification mail as an updated street
( The reason I discovered this missed update is that in a ”competing street hunting application” I was suddenly not at 100% for Solna anymore)
This is probably the same thing that I mentioned in Mismatch nested and parent city, then I thought it was somehow connected to parent/nested cities, but this one, Solna kommun, is not nested, so there seems to be something else involved
I suspect it was just a data lag between OSM & the Overpass service I use. Edits in OSM are not immediately available via Overpass - it can lag by various amounts of time. Right now it’s serving up data from 2023-09-02T20:19:00Z
I checked the Overpass query to see what it’s returning now & I see that section of Ways is included in the response. I queued up a sync for the city, so that should fix it up. It’s just going to take a while, because I’m currently running code to add the length to each street record.
Ah, the code that handles Overpass failures wasn’t resilient enough. I’m updating it so it will clear out that date as the last step, if it wasn’t successful. This will place it in the front of the queue for the next update attempt, so hopefully won’t be too confusing to people if they happen to see it change to blank.
(Note to self: see you back here in like a month when you’ve received too many emails and need to change this behavior again)
Hm, I’ve now received some more notifications, but not all, seens to be a very long lag between update finished and notification sent. And the one that was spinning is now ok, maybe the update takes a lot longer time than I thought
It’s tough to say for certain … Some take less than a minute, some take hours.
If an update runs and nothing is changed, then no notification is sent.
Let’s see how the next update goes…
Update: Merged this into the other thread on the same topic. I’m able to see that the Cambridge update started … the Overpass server that responded had old data … and it properly re-queued the city … but the next time around, with the same issue, it did not properly reset its sync date and I can’t see how/why that could happen
@JamesChevalier, I’ve submitted several ‘sync’ requests over the last few days that have immediately failed and then say that a city can only be updated once per day.
Just updated a few cities to check, they didn’t immediately update the date but when I checked back after a few minutes and refreshed the page they had an updated sync date of today. So it’s not a global issue…
I did have some new city/country additions going around the same time, which definitely would have blocked further sync requests. Those are all done, though, so let me know if it works now.
If it doesn’t, please share the city link.
I’ll need more details and maybe some links to look into it further … Technically the updates seem to be working as expected - the dates are updated, their corresponding jobs are completing successfully, etc.
I am seeing tons of issues with at least one of the Overpass servers I connect to, but all of those should self-resolve themselves via retries … and if things weren’t self-resolving, I’d expect to see lots of lingering background jobs sitting around - which I’m not seeing.
@JamesChevalier
Arlington fails immediately after requesting a sync update (many daily requests submitted recently) and says it can only be updated once a day. Sometimes the date changes, sometimes it doesn’t… but I never get an update notification/email because the update never completes even on the times when the date changes because it has failed. Here’s a street that hasn’t updated: Way: Walsh Way (8653835) | OpenStreetMap
For future requests, if a sync request fails, is there a way to be able to re-request an update (not wait a day) and/or not lose 1 of our 5 daily sync tokens?
This is already the case. In these scenarios neither the city limit nor the user limit is touched:
A) The background queue is overloaded
error message “ There are too many processing jobs running, please try again later”
B) The 5 sync limit has been met
error message “ You may only sync 5 cities per day.”
C) The city has already been synced that day
error message “ A city may only be synced once per day.”
D) The city is currently syncing
error message “ City is already syncing”
Interesting… I don’t see a path through this code where it would be possible to update the date and also not run the sync. The city’s last sync date is touched inside the code that is doing the update, so the only way for it to change is for the update to run.
Yeah, failed/incomplete updates as well as updates without changes … no emails will be sent in those cases.
I am noticing a black hole that’s definitely being traveled lately…
This can result in the sync request returning the successful “ Doin’ it” message, but the overall sync not running. I’m decently certain this is the issue you’re experiencing, because I log some data for myself to keep track of this & it looks like the last truly successful syncs for Arlington were on February 4th, 14th, and 15th.
These Overpass servers aren’t terribly reliable, so it can be common to get into a situation where the data is too old to use. I think I can work around this issue by pushing a new sync request if this scenario is met. Eventually the sync requests will hit a usable server (it randomly chooses from a collection) and the sync will complete.
I’m swamped busy with some projects, but I’ll get this change in for the next release. I can’t guarantee when that will be, but I’ll have to release some new code for these projects soon - definitely before March rolls around.
Thanks for staying on this issue. The city update code is pretty tough, so the system kind of has to accept and handle failure - but working through your reports, and seeing things through your use is showing some glaring usability issues that should be cleaned up. There’s currently a lot of guessing as to the state of the process, and I think I can do things to help clarify that.
Let’s see how the next update goes…
and I can’t see how/why that could happen 
There are too many processing jobs running, please try again later”
Doin’ it” message, but the overall sync not running. I’m decently certain this is the issue you’re experiencing, because I log some data for myself to keep track of this & it looks like the last truly successful syncs for Arlington were on February 4th, 14th, and 15th.