Created attachment 1486508 [details]
"impossible" last seen active output
Description of problem:
Observed the following string `oc describe` Conditions output: container runtime is down,PLEG is not healthy: pleg was last seen active 2562047h47m16.854775807s ago; threshold is 3m0s,network state unknown
2562047h equates to hundreds of years.
Version-Release number of selected component (if applicable):
Consistent in this particular environment for the NotReady node in question.
The node is running cri-o-1.11.1-2.rhaos3.11.git1759204.el7.x86_64.
Yes, that is a known log readability issue.
Basically is it a unit64 underflow due to the last relist time being 0. It indicates that PLEG has not succeeded, ever, since the kubelet process started.
If you see this when just starting the kubelet, it is not a real error.
If you see this after the kubelet has been up for a while, typically indicates that the runtime is not running or unresponsive.
Actually, it isn't an underflow. It is just the current time - 0 as a duration:
No time like the present to fix this. Upstream PR:
I just would like to note again (also see comment #1) this bug fix is purely cosmetic and does not point to an error with the node or container-runtime
The following log is a result time being 0 since the relist time, when the node process is first started.
PLEG is not healthy: pleg was last seen active 2562047h47m16.854775807s ago
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.