Configuration

• Startup Arguments: http://node{1...4}:9000/data/rustfs{0...3}
  (This implies a 4-node cluster, with 4 drives/volumes per node.)

Steps to Reproduce

1. Simulate Failure: node2 experienced an outage and was taken offline.
2. Write While Degraded: While the cluster was in a degraded state (with node2 offline), I successfully uploaded a new object, myname.zip.
3. Recovery: node2 was repaired and brought back online, rejoining the cluster.
4. Trigger Read: I then performed a GET request to download the myname.zip object.

Expected Behavior

My expectation is that reading the object (myname.zip) would trigger a self-heal process. The system should detect that node2 is back online but missing the newly written shards. It should then reconstruct the missing data shards for myname.zip and populate (write) them to the appropriate drives on node2 (i.e., /data/rustfs{0...3}).

Actual Result

The data shards for myname.zip are still missing from all drives on node2. The read operation succeeded (likely by reading from the other nodes), but it did not trigger the expected data reconstruction and healing process for the newly rejoined node.