Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-2705

High throughput cache bound many threaded workloads can experience long latencies

    XMLWordPrintableJSON

Details

    • Storage 2017-07-10, Storage 2017-07-31, Storage 2017-10-02, Storage 2017-10-23, Storage 2017-12-18, Storage 2018-01-01
    • 10

    Description

      In workloads with:

      • Many more threads than CPU cores
      • Enough cache pressure that the cache remains 100% full
      • Application threads are required to contribute to eviction
      • Transactions span multiple operations

      Some threads can be tasked with doing more eviction than is fair which leads to occasional operations with longer than expected latencies.

      We have gone backwards and forwards on the best approach for sharing eviction when there is lots of cache pressure and many threads.

      WT-2702 was needed to fix the code, so memory usage can't grow above 100%.
      WT-2114 was a change to improve the fairness of the code to avoid long latency operations.

      Attachments

        Issue Links

          Activity

            People

              sulabh.mahajan@mongodb.com Sulabh Mahajan
              alexander.gorrod@mongodb.com Alexander Gorrod
              Votes:
              0 Vote for this issue
              Watchers:
              14 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: