DeleteNodes checks if the node group is below
its min size. If it is and a scale up fails,
CA is unable to clean up the instances
with errors, which causes CA ending up in
an error loop. Using ForceDeleteNodes instead,
so the min size won't be validated when
removing failed instances.
The --drain-priority-config flag was only parsed if isFlagPassed()
returned true for it. However, isFlagPassed() would actually silently
never work. The implementation relied on walking the flags parsed by
the standard Go "flag" pkg. This seems like it would work since CA
defines its flags using the standard "flag" pkg. However, the flags
are then actually parsed and processed by the "github.com/spf13/pflag"
pkg, so isFlagPassed() could never see them.
This commit replaces removes isFlagPassed() and replaces the calls
with a pkg-provided pflag.CommandLine.Changed(). Unit tests are added
to verify that the flag is correctly parsed after this change.
* Update default value for scaleDownDelayAfterDelete
Setting defaut value for scaleDownDelayAfterDelete to be scanInterval
instead of 0.
* Revert the change and fix the flag description
Karol