[HDDS-3228] Enforce Exponential backoff retry policy for raft client in cases of system overload

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 0.5.0
• Fix Version/s: 0.5.0
• Component/s: Ozone Client
• Target Version/s: 1.0.0
• Assignee: Lokesh Jain
• Reporter: Shashikant Banerjee
• Created: 18/Mar/20 11:31
• Updated: 15/Apr/20 14:32
• Resolved: 15/Apr/20 14:32

Description

Currently, the retry policy enforced by ozone client on Raft client is to retry multiple times with fixed sleep interval of 1s. With system under heavy load, when server starts throwing server busy/resource unavailable exception, ideal policy would be to retry with exponential back off. This Jira aims to address this.

Issue Links

• is duplicated by: HDDS-3350 Ozone Retry Policy Improvements