The last two items were expected, but the first three are more recent (built on 20100102), which indicates that we're actually creating the issue each time we run the query fixing bug 499421. Indeed, that is the case. To fix bug 499421 when it occurs (only once over the holidays on the 1st) we're manually setting the date_started to null, but not resetting the job status.
I'll (1) update the query to be used in bug 499421, and (2) redo the query to fix the above builds so that the date_started is set more accurately.
The above query to check for the problem builds was run on production identifying the following builds:
https:/ /edge.launchpad .net/~n- muench/ +archive/ gnu-programs/ +build/ 1424385 /edge.launchpad .net/~n- muench/ +archive/ gnu-programs/ +build/ 1424386 /edge.launchpad .net/~n- muench/ +archive/ gnu-programs/ +build/ 1424387 /edge.launchpad .net/~pgquiles/ +archive/ ppa/+build/ 1401950 /edge.launchpad .net/~pgquiles/ +archive/ ppa/+build/ 1401929
https:/
https:/
https:/
https:/
The last two items were expected, but the first three are more recent (built on 20100102), which indicates that we're actually creating the issue each time we run the query fixing bug 499421. Indeed, that is the case. To fix bug 499421 when it occurs (only once over the holidays on the 1st) we're manually setting the date_started to null, but not resetting the job status.
I'll (1) update the query to be used in bug 499421, and (2) redo the query to fix the above builds so that the date_started is set more accurately.