"cannot get controller" error log in tidb-controller-manager logs

[DanielZhangQD]

Bug Report

What version of Kubernetes are you using?

1.12
What version of TiDB Operator are you using?

v3.0.5

What did you expect to see?
No such error log
What did you see instead?
The below logs are printed repeatedly:

E1230 03:44:54.263402       1 controller_utils.go:415] cannot get controller dan5/dan5-backup-s3-12191761
E1230 03:44:54.263497       1 controller_utils.go:415] cannot get controller br-test/demo1-backup-s3
E1230 03:44:54.264677       1 controller_utils.go:415] cannot get controller dan5/dan5-backup-schedule-s3-1345-2019-12-20t06-44-00
E1230 03:44:54.265805       1 controller_utils.go:415] cannot get controller dan5/dan5-backup-s3-12041551

[DanielZhangQD]

Root Cause:
TiDB Initializer watches all the jobs in the k8s cluster and tries to get the owners of the jobs from the TidbInitializer lister, however, for the jobs not created by TidbInitializer CR, the get will fail and the above logs will be printed.
Solution:
Update the handlers for job events, only to handle the jobs that are created by TidbInitializer CR.

[aylei]

1. IMO, no controller or the controller is not TidbInitializer are both normal cases, these objects could be simply ignored and no error log needed
2. adjust release CI script for release-1.0 branch #1320 may help reduce the number of objects that not controlled by TidbInitializer we see by only watching objects with the managed-by: tidb-operator label

[Yisaer]

Besides, I think the error logs should be more detailed.