[HDDS-8171] Replicate commands can be sent to dead maintenance modes if the same index is being decommissioned

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.4.0
• Labels: pull-request-available
• Assignee: Stephen O'Donnell
• Reporter: Jyotirmoy Sinha
• Created: 15/Mar/23 17:24
• Updated: 19/Mar/23 06:30
• Resolved: 19/Mar/23 06:30

Description

In a scenario where an EC container has two copies of the same replica index:

1. A maintenance replica index, where the maintenance node is dead
2. A copy which is decommissioning

The replicate commands to resolve the decommissioning index can get sent to the dead maintenance node repeatedly, and then the decommission becomes stuck.

This is because the replicas which were filtered previously are not passed into the method used to resolve the decommission and maintenance indexes.

Issue Links

• links to: GitHub Pull Request #4412