Uploaded image for project: 'Ambari'
  1. Ambari
  2. AMBARI-17510

LogSearch REST Integration component can cause performance issues

VotersWatch issueWatchersCreate sub-taskLinkCloneUpdate Comment AuthorReplace String in CommentUpdate Comment VisibilityDelete Comments
    XMLWordPrintableJSON

Details

    • Bug
    • Status: Resolved
    • Critical
    • Resolution: Fixed
    • 2.4.0
    • 2.4.0
    • ambari-server
    • None

    Description

      In larger clusters, or clusters that are long-running, some performance-related problems have been uncovered that occur when LogSearch is deployed in an Ambari cluster.

      In particular, if the LogSearch Server itself is performing slowly, this can cause a degradation of service in the Ambari REST API, since the LogSearch integration components make direct REST calls to obtain the Logging data on behalf of the Ambari UI.

      This problem is caused by two issues:

      1. The REST requests occur on the Ambari REST request thread. This can be problematic if a given request hangs. This can cause the Ambari REST API, and the Ambari server itself, to become unresponsive.

      2. The Ambari REST integration layer makes HTTP calls to the LogSearch service to obtain Logging metadata, that is then attached to the HostComponent resource in Ambari. These HTTP calls are not currently configured with a connect or read timeout, which can cause a failing call to wait indefinitely, which again can cause Ambari to become unresponsive.

      I'm working on a fix for this, and will be submitting a patch shortly.

      Attachments

        1. AMBARI-17510.patch.2
          30 kB
          Bob Nettleton

        Activity

          This comment will be Viewable by All Users Viewable by All Users
          Cancel

          People

            rnettleton Bob Nettleton
            rnettleton Bob Nettleton
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Issue deployment