Uploaded image for project: 'Indy Node'
  1. Indy Node
  2. INDY-1866

Possible memory leak during catchup

    XMLWordPrintable

    Details

    • Type: Task
    • Status: Complete
    • Priority: Medium
    • Resolution: Done
    • Affects Version/s: None
    • Fix Version/s: 1.6.83
    • Component/s: None
    • Labels:
      None

      Description

      During long sustained load test one of nodes was restarted, successfully completed short catchup and continued ordering. However after that all other nodes increased memory consumption and number of live objects, but number of requests in their queues (which is major contributor to number of live objects) was not increased. This increase in number of live objects seems constant. There is probability that node has memory leak related to catchup and/or other activity related to return of some node to consensus.

      PoA

      • run load test in docker with on of nodes continuosly restarted
      • check memory consumption and number of live objects on all nodes except one being restarted
      • if there is increase in memory consumption try to identify its source (logging top 50 types of live objects can help here)
      • if there is an identified leak either fix it or create issue dedicated to fix (if fix would take more than 1 day)

        Attachments

          Activity

            People

            Assignee:
            sergey.khoroshavin Sergey Khoroshavin
            Reporter:
            sergey.khoroshavin Sergey Khoroshavin
            Watchers:
            Sergey Khoroshavin
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Dates

              Created:
              Updated:
              Resolved: