Description of problem: After upgrading one of the nodes of a 2.0-all cluster to 2.1, the peer gets rejected with volume checksum mismatch. Version-Release number of selected component (if applicable): RHS-2.1 built from source. How reproducible: Steps to Reproduce: 1. Create a cluster of 2 nodes, both operating in version 2.0. 2. Create a volume with bricks on both the nodes. 3. Start the volume. 4. Mount the volume from a 2.0 client. 5. Upgrade one of the servers to 2.1. 6. Restart glusterd on this 2.1 node. 7. The node gets rejected with volume cksum mismatch. Actual results: Volume cksum computed on the two nodes differ. Expected results: Volume cksum must be the same on both the nodes. Additional info:
The fix is available in glusterfs-v3.4.0.33rhs, hence moving the state of the bug to ON_QA.
This is a sanity only bug. Ref: https://mojo.redhat.com/docs/DOC-20570 - Upgrade tests.