Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-46925

XDCR - large memory footprint for pipeline supervisor with large number of replications

    XMLWordPrintable

Details

    • Bug
    • Resolution: Fixed
    • Major
    • 7.1.0
    • CheshireCat.Next
    • XDCR
    • None
    • 1

    Description

      To investigate further: When creating a 12x12 mesh between two clusters, XDCR memory usage seems to balloon up (7G in RSS) without actually any data transfer

      The test:

      1. Sets up 2 1-node clusters
      2. Create 12 buckets on each cluster, 100MB ramQuota
      3. For each bucket, create a replication to all target buckets, creating a mesh
      4. Replication has: 20000 statsInterval and 600 checkpointInterval
      5. The following internal settings are set:

      setInternalSettingsWithoutImport "C1" "TopologyChangeCheckInterval=60"
      setInternalSettingsWithoutImport "C2" "TopologyChangeCheckInterval=60"
      setInternalSettingsWithoutImport "C1" "RefreshRemoteClusterRefInterval=60"
      setInternalSettingsWithoutImport "C2" "RefreshRemoteClusterRefInterval=60"
      setInternalSettingsWithoutImport "C1" "ReplSpecCheckInterval=60"
      setInternalSettingsWithoutImport "C2" "ReplSpecCheckInterval=60" 
      

      Attachments

        1. 06152021_node0.log
          26.66 MB
          Neil Huang
        2. heap0.pprof.pdf
          184 kB
          Neil Huang

        Issue Links

          For Gerrit Dashboard: MB-46925
          # Subject Branch Project Status CR V

          Activity

            People

              neil.huang Neil Huang
              neil.huang Neil Huang
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Gerrit Reviews

                  There are no open Gerrit changes

                  PagerDuty