Comment 3 for bug 499095

I'm just trying to piece together the original situation, after struggling to see how the build could have been re-dispatched after being set as superseded.

The original issue according to the internal irc log that: the slave scanner is stuck with "Scanning failed with: 'NoneType' object has no attribute 'slaveStatus'".

The cause was this corrupt build, which at the time thought it had already started building (job.date_started is not null) but it didn't have an associated builder (bq.builder is null) as shown by a count of 1 from the query https://pastebin.canonical.com/25913/

The fix taken on that date was not actually paste 25907 above, but https://pastebin.canonical.com/25914/, which simply set the job.date_started back to null for this build. So it's expected that it would be dispatched again.

OK, now to figure out why it's stalling on hassium...