1 Reply Latest reply on Dec 13, 2012 4:31 PM by dsbalcau

    NetApp cifs operations

    tomaddox

      I'm trying to troubleshoot a potential performance issue with a NetApp, and I have a question about how Storage Manager displays CIFS operations. If I look at CIFS Total Ops/second, I see a peak of around 5000 at 7:00 AM (for example), while the Ops Per Sec counter is around 8000 at the same time. CIFS Writes/sec at that time are below 1.00. Reads/sec are a flat 0.00, as are Others/sec. I guess I have two questions:

       

      1) What is the difference between Ops Per Sec and CIFS Total Ops/sec (both are listed under Performance->CIFS)?

      2) How can I tell what operations are actually taking place, since apparently they're not read, write, or other?

       

      TIA,

       

      Tom

        • Re: NetApp cifs operations
          dsbalcau

          Hi Tom,

           

          Thanks for your question. I think you may have stumbled across a bug. Can you please file a support ticket for this case?

           

          So this is the difference between those two metrics:

           

          We are querying two different objects in the NetApp API:

          For Ops Per Sec - Object: cifs_ops  Metric: total_ops       8B      RAW     NONE    Total number of SMB operations since filer was started  CM_DIAG       

           

           

          For CIFS Total Ops/sec - Object: cifs Metric: cifs_ops        8B      RATE    PER_SEC         Total number of CIFS operations per second      CM_BASIC      

           

          I think the bug is that we are treating both as pure counters, when one is actually just a pure counter and the other is a rate. In this case, I would trust Ops Per Sec and not CIFS Total Ops/sec.

           

          Also, it looks like there may be a problem with the way we calculate the "others" (i.e. non-Read/Write) metric. Like I said,to verify this, please submit a ticket and we can verify from your diagnostics exactly what the root cause is. Sorry for the inconvenience.

           

          Regarding what may be causing the spike in CIFS ops (regardless of the number) in the "other" category, it could be either antivirus or File Analysis if you have that enabled in Storage Manager.

          See this post for a similar discussion: cifs operations | Netapp | toasters