Many builds stuck in "triggered" state

[gatesphere]

I have a script that runs for the leo-editor project requesting a nightly build. Every build since July 24th has been stuck in the "Triggered" state. Is there any way to cancel/delete these builds? I have disabled the nightly script that triggers these builds until I can find a solution to this problem.

[mitechie]

I've got this with a fresh project I setup 'charmworld'. Each build is stuck in triggered state for the last 3 days. Since it's in 'triggered' there's no failure message, logging, or other utility I can see to debug the issue.

[ericholscher]

It's almost certainly our bazaar support being broken. Nobody on the core team, and very few of our users use bzr, so it's support is half-baked at best.

[ericholscher]

Likely #349

[gatesphere]

:-/

[gatesphere]

Our team "fixed" this by setting up a nightly git mirror here on github. It pulls from bzr and pushes to git nightly, and then fires off the curl POST to request another build.

That said, we have 2 weeks of docs still stuck in "Triggered" state.

[cjwatson]

@gatesphere: Is it better now? I fixed #349 and my project is working again.

[gatesphere]

Still have several builds in "Triggered" state, but that doesn't matter. The recent builds have been working successfully, but they're reading from our nightly git mirror. I'm unlikely to try with bzr again, as the git mirror serves a dual purpose anyway.

Thanks for checking in, though.

[ericholscher]

Closing this for now. Thanks @cjwatson!

[kitelife]

@ericholscher my project is also stucked in “triggered” state

[kitelife]

@ericholscher The source code of this project is hosted on github，and i add “readthedocs” hook for this project。This project is here,Thank you!

[codedstructure]

I had exactly the same issue with a (Mercurial) project (pylibftdi) hosted on bitbucket, and 'resolved' it in the same way by pointing to a github mirror - so it seems it isn't just bzr which is having / had issues.

[cjwatson]

Perhaps there should be separate issues for the git and Mercurial problems, since the bzr problems were definitely traceable to a bug specifically in rtfd's bzr code.

Any bug that causes the import of a project to crash will result in this symptom, so separate instances of it should really be separate issues in the same way that different crashes in any project should be separate bugs.

[codedstructure]

Thanks, opened new issue as #547