nodedeamon pods alternating between Running and CrashLoopBackOff states
In CODECO installation the nodedaemon pods by SWM seem to alternate between Running and CrashLoopBackOff states, restarting all the time:
kube-system nodedaemon-4mvxh 1/1 Running 15 (6m46s ago) 75m
kube-system nodedaemon-jhzt4 1/1 Running 15 (6m5s ago) 75m
kube-system nodedaemon-v7s6h 1/1 Running 15 (6m15s ago) 75mJust a minute after:
kube-system nodedaemon-4mvxh 0/1 CrashLoopBackOff 15 (35s ago) 76m
kube-system nodedaemon-jhzt4 0/1 Error 15 (6m40s ago) 76m
kube-system nodedaemon-v7s6h 0/1 CrashLoopBackOff 15 (13s ago) 76mActivity
added Bug label
assigned to @mueller9001 and @jkgesswein
The error is uncritical, as the nodedaemon does all necessary things before it runs into this error, which leads to a restart.
The error is caused by the final select() call of the nodedaemon main() function. The log entry is
fatal error: all goroutines are asleep - deadlock! goroutine 1 [select (no cases)]: main.main() /builds/itp_cloud_research/qos-scheduler/scheduler/cmd/node-daemon/node-daemon.go:48 +0x1c5mentioned in task eclipse-research-labs/codeco-project/use-cases/p5-amr-manufacturing#7
marked this issue as related to eclipse-research-labs/codeco-project/use-cases/p5-amr-manufacturing#7
marked this issue as related to eclipse-research-labs/codeco-project/use-cases/p3-mec-enablement#1
marked this issue as related to eclipse-research-labs/codeco-project/use-cases/p2-vehiculardt/v2xservices/btp#1
