Fix non-zero reference count for replica shard files in CompositeDirectory

Describe the bug

When a new file is created in CompositeDirectory, its reference count is increased by 1 in the FileCache. This count is normally decreased when the file is uploaded to remote storage. However, for replica shards, a segment_n file is created but never uploaded, leaving its reference count at 1. This causes problems during index deletion, as some files can't be deleted due to non-zero references.

We need to evaluate two options:

• Prevent file creation on replica shards for CompositeDirectory.
• If files must be created on replica shards, ensure they don't have non-zero references in the FileCache.
  This issue affects clean index deletion and may lead to resource leaks if not addressed.

Meta Issue: https://github.com/opensearch-project/OpenSearch/issues/13149

Related component

No response

To Reproduce

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Expected behavior

.

Additional Details

Plugins
Please list all plugins currently enabled.

Screenshots
If applicable, add screenshots to help explain your problem.

Host/Environment (please complete the following information):

• OS: [e.g. iOS]
• Version [e.g. 22]

Additional context
Add any other context about the problem here.

Start by locating CompositeDirectory and FileCache, then trace how replica-shard segment_n files are created, uploaded, and released. Compare the two proposed handling options and verify the chosen behavior against index deletion; done means replica files no longer retain non-zero references and clean deletion succeeds, with regression coverage for the failure.