Red Hat Bugzilla – Bug 1357786
After update build runPolicy from serial to paralle,the first parallel build in queue will run in serial
Last modified: 2017-03-08 13:26 EST
Description of problem: Trigger one more serial builds, then change builds runPolicy to Parallel, trigger 3 parallel builds.The first parallel build in quene will run in serial after all serial builds complete. Version-Release number of selected component (if applicable): openshift v3.3.0.6 kubernetes v1.3.0+57fb9ac How reproducible: always Steps to Reproduce: 1.Trigger one more serial builds. 2.Change bc's runPolicy to Parallel, trigger 3 parallel builds 3.Check builds Actual results: After the serial builds complete, the first parallel build will run serially,others keeping "new" status, after the first parallel build complete, the remain two builds will run parallel. Expected results: After the serial builds complete, all the parallel builds including the first one should begain to run parallely. Additional info:
Hi, Ben, any pr on this bug ? or you don't think it's a bug ?
Sorry, I swear i linked the PR here but i guess not...it was a valid bug, i've fixed it here: https://github.com/openshift/origin/pull/9969
(In reply to Ben Parees from comment #2) > Sorry, I swear i linked the PR here but i guess not...it was a valid bug, > i've fixed it here: > https://github.com/openshift/origin/pull/9969 Hi, Ben , considering this scenario, there are 3 serial + 3 parallel + 3 serial + 3 parallel builds in the queue, after the first 3 serial builds completed one by one, there are total 6 parallel builds in the queue , but should we run the 6 builds parallel or just run the following 3 parallel builds, then 3 serial builds, then the last parallel 3 builds ?
the first 3 serial will run serially, then all 6 parallel simultaneously, then the second 3 serial will run serially. i had some debate about that behavior, but it would be significantly more complicated to do it as 3, 3, 3, 3, and i don't know that running all 6 is wrong. (plus it's not likely to be a real use case for users).
(In reply to Ben Parees from comment #4) > the first 3 serial will run serially, then all 6 parallel simultaneously, > then the second 3 serial will run serially. > > i had some debate about that behavior, but it would be significantly more > complicated to do it as 3, 3, 3, 3, and i don't know that running all 6 is > wrong. (plus it's not likely to be a real use case for users). Ok, if we designed for this , mark this verified.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2016:1933