[CASSANDRA-13074] DynamicEndpointSnitch frequently no-ops through early exit in multi-datacenter situations

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 2.2.9, 3.0.11, 3.10
• Component/s: Legacy/Coordination
• Severity: Normal
• Assignee: Joel Knighton
• Reporter: Joel Knighton

Authors

Joel Knighton

Reviewers

T Jake Luciani

• Created: 22/Dec/16 18:30
• Updated: 16/Apr/19 09:30
• Resolved: 03/Jan/17 18:07

Description

The DynamicEndpointSnitch attempts to use timings from nodes to route reads to better performing nodes.

In a multi-datacenter situation, timings will likely be empty for nodes outside of the local datacenter, as you'll frequently only be doing local_quorum reads (or a lower consistency level). In this case, the DES exits early and returns the subsnitch ordering. This means poorly performing replicas will never be avoided, no matter how degraded they are.