-
Bug
-
Resolution: Done-Errata
-
Critical
-
None
-
4.16.0, 4.17.0
-
None
-
No
-
Approved
-
False
-
-
N/A
-
Release Note Not Required
-
Done
Description of problem:
If infrastructure or machine provisioning is slow, the installer may wait several minutes before declaring provisioning successful due to the exponential backoff. For instance, if dns resolution from load balancers is slow to propogate and we
Version-Release number of selected component (if applicable):
How reproducible:
Sometimes, it depends on provisioning being slow.
Steps to Reproduce:
1. Provision a cluster in an environment that has slow dns resolution (unclear how to set this up) 2. 3.
Actual results:
The installer will only check for infrastructure or machine readiness at intervals of several minutes after a certain threshold (say 10 minutes).
Expected results:
Installer should just check regularly, e.g. every 15 seconds.
Additional info:
It may not be possible to definitively test this. We may want to just check ci logs for an improvement in provisioning time and check for lack of regressions.
- blocks
-
OCPBUGS-35383 capi installer reconciliation takes longer than necessary
- Closed
- is cloned by
-
OCPBUGS-35383 capi installer reconciliation takes longer than necessary
- Closed
- relates to
-
OCPBUGS-38657 capv session timeout
- Verified
- links to
-
RHEA-2024:3718 OpenShift Container Platform 4.17.z bug fix update