Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-2594

Optimize eviction for in-memory workloads

    • Type: Icon: Improvement Improvement
    • Resolution: Won't Fix
    • Priority: Icon: Major - P3 Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • None

      WiredTiger's eviction heuristics were designed for workloads that can overflow a cache by evicting pages from cache to make space.

      In-memory workloads only use "eviction" for garbage collection. We should come up with a realistic in-memory workload that generates old versions but doesn't have more data than the memory size. Then measure how effective eviction is at identifying candidates that will all space to be freed. We will need to change how we prioritize pages when running in-memory.

            Assignee:
            backlog-server-storage-engines [DO NOT USE] Backlog - Storage Engines Team
            Reporter:
            michael.cahill@mongodb.com Michael Cahill (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated:
              Resolved: