ClientRetryPolicy: Fixes behavior to Meta-data write operations in multimaster accounts

[NaluTripician]

Pull Request Template

Description

In the issue, for multimaster accounts where the LimitToEndpoint flag is set to true and the endpoint is specified to the non-hub region, all metadata write operations will fail because we are trying to perform a write on a non-bub region that does not have access to the master partition. The failure will be a 403.3 error and the current way the ClientRetryPolicy is set up, these metadata write operations will not retry as the enableEndpointDiscovery flag is set to false when LimitToEndpoint is true. In the LocationCache. By changing the behavior of the 'ClientRetryPolicy' to retry metadata write opeations and LocationCache to check to see if it is a retried metadata write operation and then to reroute to the hub region, this problem can be mitigated.

Type of change

Please delete options that are not relevant.

• [] Bug fix (non-breaking change which fixes an issue)
• [] New feature (non-breaking change which adds functionality)

Closing issues

To automatically close an issue: closes #3408

[ealsur]

Please add tests to cover any changes to the files and behaviors and see the other comments, good starting point!