-
Bug
-
Resolution: Won't Do
-
Normal
-
None
-
4.12
-
Important
-
No
-
3
-
Sprint 241 - Pipeline&Delivery, Sprint 242 - Pipeline&Delivery, Sprint 243 - Pipeline&Delivery
-
3
-
False
-
Description of problem:
Customer is using the following crash command to test resilience of their Openshift cluster nodes "echo c > /proc/sysrq-trigger". When this command is executed it should crash the node and if kkdump is configured should also generate a dump. When we are testing this command on any Openshift cluster running on Azure cloud, the nodes are going into notReady but hung there, nothing happens and the node does not restart. For you to regain access to the node you must use the Azure UI to force restart the node.
Version-Release number of selected component (if applicable):
4.10.x 4.11, 4.12.x
How reproducible:
every time
Steps to Reproduce:
1. Access the node, via oc debug node/<NodeName> or SSH 2. switch to root user and or chroot /host 3. type: echo c > /proc/sysrq-trigger
Actual results:
your session will hung and the node will stay into notReady until you force restart it using the Azure UI
Expected results:
The node should crash, go into not Ready then restart and become Ready again
Additional info:
This issue seems to be isolate to Openshift clusters running Azure cloud. We have tested on vmWare and AWS clusters and it does work as expected.