Disallow setting both balance.index and balance.shard factors to 0

[hlcianfagna]

Problem Statement

Performing the following configuration is allowed:

set global persistent "cluster.routing.allocation.balance.index" = '0.00f';
set global persistent "cluster.routing.allocation.balance.shard" = '0.00f';

However after restarting the nodes the cluster will no longer start and the following entries will appear in the logs:
failed to apply updated cluster state
Balance factors must sum to a value > 0 but was: 0.0

Possible Solutions

• Reject changing one of these settings to zero if the other one is already zero

Considered Alternatives

No response