Backport PR 36812 from upstream to v1.4.0-rc1

[ksurent]

There was a bug in the upstream where Jobs and Pods created as a result of a ScheduledJob run would have names that weren't unique enough, which prevented the ScheduledJob from completing.
It was fixed in kubernetes/kubernetes#36812.

Without that fix it's really hard to reliably run recurring tasks in Origin so I propose it should be backported to Origin v1.4.0+rc1 and carried over to later v1.4.* releases.

Version

oc v1.4.0-rc1+b4e0954
kubernetes v1.4.0+776c994
features: Basic-Auth GSSAPI Kerberos SPNEGO

Server https://localhost:8443
openshift v1.4.0-rc1+b4e0954
kubernetes v1.4.0+776c994

Steps To Reproduce

1. Create a minutely ScheduledJob
2. Wait for some time

Current Result

The job stops running after some time.

Expected Result

The job should keep running.

Additional Information

Logs from origin-master:

1 event.go:217] Event(api.ObjectReference{Kind:"ScheduledJob", Namespace:”ksurent-jobs", Name:"hellocron", UID:"aef3a4f1-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"582233", FieldPath:""}): type: 'Normal' reason: 'SuccessfulCreate' Created job hellocron-1706264388
1 event.go:217] Event(api.ObjectReference{Kind:"Job", Namespace:”ksurent-jobs", Name:"hellocron-1706264388", UID:"cac57d81-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"582319", FieldPath:""}): type: 'Normal' reason: 'SuccessfulCreate' Created pod: hellocron-1706264388-zb0vb
1 event.go:217] Event(api.ObjectReference{Kind:"Job", Namespace:”ksurent-jobs", Name:"hellocron-1706264388", UID:"cac57d81-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"582319", FieldPath:""}): type: 'Normal' reason: 'SuccessfulCreate' Created pod: hellocron-1706264388-v9jle
1 event.go:217] Event(api.ObjectReference{Kind:"Job", Namespace:”ksurent-jobs", Name:"hellocron-1706264388", UID:"cac57d81-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"582319", FieldPath:""}): type: 'Normal' reason: 'SuccessfulCreate' Created pod: hellocron-1706264388-yflxn
1 event.go:217] Event(api.ObjectReference{Kind:"ScheduledJob", Namespace:”ksurent-jobs", Name:"hellocron", UID:"aef3a4f1-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"582370", FieldPath:""}): type: 'Normal' reason: 'SawCompletedJob' Saw completed job: hellocron-1706264388

... 1 hour passes ...

1 event.go:217] Event(api.ObjectReference{Kind:"ScheduledJob", Namespace:”ksurent-jobs", Name:"hellocron", UID:"aef3a4f1-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"585993", FieldPath:""}): type: 'Warning' reason: 'FailedCreate' Error creating job: jobs.batch "hellocron-1706264388" already exists
1 event.go:217] Event(api.ObjectReference{Kind:"ScheduledJob", Namespace:”ksurent-jobs", Name:"hellocron", UID:"aef3a4f1-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"585993", FieldPath:""}): type: 'Warning' reason: 'FailedCreate' Error creating job: jobs.batch "hellocron-1706264388" already exists
1 event.go:217] Event(api.ObjectReference{Kind:"ScheduledJob", Namespace:”ksurent-jobs", Name:"hellocron", UID:"aef3a4f1-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"585993", FieldPath:""}): type: 'Warning' reason: 'FailedCreate' Error creating job: jobs.batch "hellocron-1706264388" already exists
1 event.go:217] Event(api.ObjectReference{Kind:"ScheduledJob", Namespace:”ksurent-jobs", Name:"hellocron", UID:"aef3a4f1-d427-11e6-a4dd-9cdc7163ec00", APIVersion:"batch", ResourceVersion:"585993", FieldPath:""}): type: 'Warning' reason: 'FailedCreate' Error creating job: jobs.batch "hellocron-1706264388" already exists

... keeps going like this ...

1 controller.go:163] Cannot determine if ksurent-jobs/hellocron needs to be started: Too many missed start times to list

[soltysh]

ScheduledJob is an alpha feature without any support, sorry :( This bug will be fixed in 1.5 where ScheduledJob is renamed to CronJob.