Fix indexer deletion in upgrade scenario

[tmaltaverne]

This PR addresses issue #1151.
There are 2 changes:

• Make sure that the fix of PR cspl-2253: indexer deletion in upgrade scenario #1041 is applied in the cluster-manager usecase
• Fix the check on the image version. The format of the image field being like splunk/splunk-operator:8.2.0, it is needed to get the version that is after the colon. The previous check was only checking if the image was starting with 8 or 9, which is always false.

[Philbow]

Discovered the same problem when triggering a helm upgrade from 2.2.1 to 2.3.0. Unfortunately the operator is not logging out the deletion of the STS.

Any plans to fix the pull request in order to merge it for the upcoming release? This is currently blocking the seamless upgrade for our Splunk customers.