I had a build that was going for 3 hours, it seemed to publish, but it still listed as building. Another was queued behind it, so I canceled it. Now it’s suck in “Canceling” and the queued build is still queued.
I have the same issue with several of my builds – except they were running since last night (12+ hours).
I can even see the build log and it shows “build finished”. Cancelling does not seem to get the builds unstuck.
Any ideas on what to do?
Same with my build, stuck for 8 hours already and no answer from docker support.
I guess creating new repository is only option now.
(Glad that it is a hobby-project image, same situation with buisness project will be unfortunate)
Well, this morning I looked and both showed up as canceled. I don’t know if there’s someone who looked into it, or if there’s some kind of automated reaper that slays all of the builds beyond a certain age.
Once that happened I was able to manually kick off a rebuild. While I’m not super annoyed or anything, it does put a little tarnish on completely automated deployment pipline 
Well, it seems to be doing it again. Am I doing something wrong? Is there some kind of rule that is assigning me to a instance with a bad actor? Is the dockerhub UI just FUBAR?
Can someone from Docker comment?
EDIT: this is kinda frustrating. I’m now paying for this service (not a lot, I realize, but I am paying), and it seems like it worked better before I opened my wallet.
EDIT2: it finally finished after 4hrs. the only part that was not cached was the last step of the build, a copy line that copies less then 20kb of files. something is seriously wrong here.