[KUDU-807] Retries are done in a tight loop when we can't resolve

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: M5
• Component/s: client
• Target Version/s: Feature Complete
• Assignee: Jean-Daniel Cryans
• Reporter: Jean-Daniel Cryans
• Created: 04/Jun/15 15:34
• Updated: 26/Feb/16 12:08
• Resolved: 09/Jun/15 16:06

Description

Saw this after trying Ricky Saltzer's code remotely against a machine running in OpenStack, the master didn't send the FQDN back to the client which it couldn't resolve:

[ERROR - org.kududb.client.AsyncKuduClient.getIP(AsyncKuduClient.java:1571)] Failed to resolve the IP of `host-10-17-81-195' in 24000ns

The returned error message shows that we looped a lot and didn't sleep much:

Exception in thread "main" org.kududb.client.NonRecoverableException: Too many attempts: KuduRpc(method=Write, tablet=null, attempt=101, DeadlineTracker(timeout=10000, elapsed=1068), null)