1 Reply Latest reply: Jan 17, 2013 1:34 AM by Srinath Menon-Oracle RSS

    One node of UCM Cluster performing very slow for webdav/DIS requests

    Prateek Mohan
      Dear All,

      One of our UCM 11gR3 nodes is taking too long to service webdav/DIS requests.

      Where first node can render a folder with say 400 subfolders, in 3-4 sec; the other one takes atleast 70 sec (and even times out at times)!
      These stats are after we eliminate our LoadBalancer/OHS effects, by directly accessing Managed Server URL.

      Web-interface and Search Performance-vise, both nodes seem to perform fine & similarly.

      The behavior is only being seen since yesterday morning. No changes have happened on the system.
      The only action done was to "Clear Statement Cache" under DataSource for both UCM managed server.

      We have already tried restarting the managed servers. First node is still doing great but second one is still in distress.
      I have verified that it is not a space/memory related issue.
      Any pointers what could be wrong!


      Regards,
      Prateek
        • 1. Re: One node of UCM Cluster performing very slow for webdav/DIS requests
          Srinath Menon-Oracle
          Hi Prateek ,

          First and best way to see what is happening with this particular node would be to set the following trace sections with full verbose tracing :

          folders*,requestaudit,systemdatabase,socketrequest

          With these tracing sections set access from DIS 2nd node and wait till the final results come .

          Once either the folders show up or operation times out navigate to View server output and refresh the section , capture the logs to see if there is any particular section that creates this bottleneck .

          Also check how good is the connection from DIS to the second node host machine ? That also might create a specific issue like this .

          Thanks,
          Srinath