Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-18597

n1ql (stale=false) : CI2 test is slower than Q2.

    XMLWordPrintable

Details

    • Untriaged
    • Unknown

    Description

      Tested with 4.5.0-1698.
      Q2=false throughput test.
      CI: 915 qps
      Q2: 1038 qps.

      Attachments

        No reviews matched the request. Check your Options in the drop-down menu of this sections header.

        Activity

          Hi Wayne, While CI2 with stale=false is lower than Q2, the numbers are >30% better than 4.1 (5005) release. So, this has actually improved significantly.
          Of course Q2(stale=false) has improved compared 4.1, albeit at a slightly better rate.

          I'll ask secondary index & forstdb teams to look at it as well.

          keshav Keshav Murthy added a comment - Hi Wayne, While CI2 with stale=false is lower than Q2, the numbers are >30% better than 4.1 (5005) release. So, this has actually improved significantly. Of course Q2(stale=false) has improved compared 4.1, albeit at a slightly better rate. I'll ask secondary index & forstdb teams to look at it as well.
          jliang John Liang added a comment - - edited

          IMO, for measuring covering index, we should use stale=ok test. This will stress out the code path in indexer and cbq more. Stale=false is mostly depends on storage write throughput, so it is not a good indication of effectiveness of covering index (read throughput)

          For stale=ok, CI2 and Q2 are about the same. But CI3 throughput is twice of Q3. This clearly shows covering index can yield better performance in typical use case (with larger row count).

          jliang John Liang added a comment - - edited IMO, for measuring covering index, we should use stale=ok test. This will stress out the code path in indexer and cbq more. Stale=false is mostly depends on storage write throughput, so it is not a good indication of effectiveness of covering index (read throughput) For stale=ok, CI2 and Q2 are about the same. But CI3 throughput is twice of Q3. This clearly shows covering index can yield better performance in typical use case (with larger row count).

          The stale=false with forestdb is in reasonable range. I'm closing this issue now as we don't intend to do additional work improve this. The work on lightweight snapshot may improve this number.

          keshav Keshav Murthy added a comment - The stale=false with forestdb is in reasonable range. I'm closing this issue now as we don't intend to do additional work improve this. The work on lightweight snapshot may improve this number.

          Bulk closing invalid bugs

          raju Raju Suravarjjala added a comment - Bulk closing invalid bugs

          People

            keshav Keshav Murthy
            wayne Wayne Siu
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Gerrit Reviews

                There are no open Gerrit changes

                PagerDuty