History syncers memory leak with Elasticsearch

XMLWordPrintable

    • Type: Problem report
    • Resolution: Won't fix
    • Priority: Trivial
    • 4.0 (plan)
    • Affects Version/s: 3.4.7
    • Component/s: Server (S)
    • None
    • Sprint 30, Sprint 31, Sprint 32
    • 1

      Hello,
      After switching to Elasticsearch history store we are getting out of memory crashes

      Log
       23297:20180320:085246.277 [file:expression.c,line:4470] zbx_malloc: out of memory. Requested 12581208 bytes.
 22750:20180320:085246.393 One child process died (PID:23297,exitcode/signal:1). Exiting ...
 22750:20180320:085248.639 syncing history data...
 22750:20180320:085249.675 syncing history data done
 22750:20180320:085249.675 syncing trend data...
 22750:20180320:085249.675 syncing trend data done
 22750:20180320:085249.675 Zabbix Server stopped.

      StartDBSyncers set to 48

      We din`t have the same issue when historical storage was in Oracle

        1. cacheUsage.jpg
          197 kB
          mzhuravlev
        2. memoryUsed.jpg
          117 kB
          Mikhail Markov

            Assignee:
            Michael Veksler
            Reporter:
            Mikhail Markov
            Team C
            Votes:
            0 Vote for this issue
            Watchers:
            8 Start watching this issue

              Created:
              Updated:
              Resolved: