Loading...

XML

Word

Printable

Description of problem:

    Stopping instances via the aws console results in instances stuck in "Stopped" and machines attempting to delete but failing due to e.g. pdbs.

Version-Release number of selected component (if applicable):

    4.14.10

How reproducible:

    100%

Steps to Reproduce:

ROSA cluster:
    1. Stop aws instances (default nodepool)
    2. Wait ~30 min until cluster-api attempts to delete the machines

Actual results:

   EC2 instances are left in "Stopped", machines are in "Deleting" state

Expected results:

    Nodepool self-repairs to expected state e.g. by restarting the stopped instances.

Additional info:

    Must gather attached