[CAMEL-10432] concurrentlinkedhashmap used by LRU cache consumes 100MB memory on 1024 core Solaris T5

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.17.1
• Fix Version/s: 2.18.0
• Component/s: camel-core

Environment

Solaris T5 1024 core machine
java 1.7.045

• Estimated Complexity: Unknown
• External issue ID: https://github.com/ben-manes/concurrentlinkedhashmap/issues/49
• Assignee: Claus Ibsen
• Reporter: Matt Baker
• Created: 01/Nov/16 15:41
• Updated: 04/Nov/16 15:59
• Resolved: 04/Nov/16 15:59

Description

Cloned from: AMQ-6492

On a Solaris T5 with 1024 cores, the amount of memory used by LRUCache which in turn is using ConcurrentLinkedHashMap from
https://github.com/ben-manes/concurrentlinkedhashmap

consumes about 20MB of memory. In my install this correlates to over 300MB of memory.

concurrentlinkedhashmap allocates memory based on CPU count, with no upper limit, so each instance of LRUCache in memory consumes more memory than is really needed.

See AMQ-6492 for screenshots (didn't want to copy/paste them here)