[GEODE-8620] Actual redundancy of -1 in restore redundancy result

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.13.0
• Fix Version/s: 1.13.1, 1.14.0
• Component/s: gfsh, management
• Labels: pull-request-available
• Assignee: Donal Evans
• Reporter: Aaron Lindsey
• Created: 16/Oct/20 00:01
• Updated: 18/Nov/20 00:37
• Resolved: 21/Oct/20 17:09

Description

Steps to reproduce:

1. Create a geode cluster with 1 locator and 2 servers.
2. Create a region of type PARTITION_REDUNDANT.
3. Put an entry into the region.
4. Trigger a restore redundancy operation via the management REST API or gfsh.
5. The result from the restore redundancy operation states that the actual redundancy for the region is -1. However, the expected redundancy at this point is 1 because there should be enough cache servers in the cluster to hold the redundant copy.
6. Stop one of the servers.
7. Trigger another restore redundancy operation via the management REST API or gfsh.
8. The result from the second restore redundancy operation again states that the actual redundancy for the region is -1. However, the region should be counted as having zero redundant copies at this point because there is only one cache server.

I encountered this issue while using the management REST API, although the same issue happens in the gfsh command. I assume fixing the gfsh command would also fix the management REST API. If not, I can break this out into two separate JIRAs.

Issue Links

• links to: GitHub Pull Request #5642