-
Bug
-
Resolution: Done-Errata
-
Undefined
-
None
-
4.19.0
-
Quality / Stability / Reliability
-
False
-
-
3
-
None
-
None
-
None
-
None
-
uShift Sprint 266
-
1
-
Done
-
Bug Fix
-
-
None
-
None
-
None
-
None
Description of problem:
MicroShift on startup creates .nodename file which holds the last hostname. If MicroShift's startup is interrupted (e.g. by reboot issued by microshift-tuned daemon), the file can be empty. If the file is empty, it'll be used on next MicroShift startup causing it to think that node name is empty string. That empty string will be stored in kubelet's certificate (CN: system:node:) which will cause API Server to reject kubelet's calls because user "system:node:" doesn't have an access to Node resource.
Version-Release number of selected component (if applicable):
How reproducible:
Once in 7 months since the low latency test was added.
Steps to Reproduce:
1. Setup MicroShift with low latency - microshift-tuned daemon should restart the host on first boot to activate new tuned profile.
2.
3.
Actual results:
Kubelet cannot register node because of the malformed user.
Expected results:
MicroShift starts successfully (aka .nodename is created atomically)
Additional info:
- links to
-
RHEA-2024:11040
Red Hat build of MicroShift 4.19.z bug fix and enhancement update