[YARN-5008] LeveldbRMStateStore database can grow substantially leading to long recovery times

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.7.3
• Fix Version/s: 2.8.0, 2.7.3, 3.0.0-alpha1
• Component/s: resourcemanager
• Target Version/s: 2.7.4
• Hadoop Flags: Reviewed
• Assignee: Jason Darrell Lowe
• Reporter: Jason Darrell Lowe
• Created: 28/Apr/16 14:30
• Updated: 06/Jan/17 01:02
• Resolved: 29/Apr/16 04:52

Description

On large clusters with high application churn the background compaction in leveldb may not be able to keep up with the write rate. This can lead to large leveldb databases that take many minutes to recover despite not having very much real data in the database to load. Most the time is spent traversing tables full of keys that have been deleted.