Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-28712

serverStatus collect extraInfo may be slow?

    XMLWordPrintableJSON

Details

    • Icon: Bug Bug
    • Resolution: Works as Designed
    • Icon: Major - P3 Major - P3
    • None
    • None
    • Performance
    • None
    • ALL

    Description

      In mongod's slow log. I got so many lines like below:
      Mon Apr 10 11:05:01.426 I COMMAND [ftdc] serverStatus was very slow:

      { after basic: 0, after asserts: 0, after connections: 0, after extra_info: 1410, after globalLock: 1410, after locks: 1410, after network: 1410, after opcounters: 1410, after opcountersRepl: 1410, after repl: 1410, after storageEngine: 1410, after tcmalloc: 1410, after wiredTiger: 1410, at end: 1420 }

      there are so many lines like above.
      from the ftdc flag, I know it is generated from the diagonal thread, but this is not important.

      What is important is that time all grows after getting extra_info. Which I found is to run mallinfo() on linux after digging into the code.

      I doubt that mallinfo may be slow in some situations because all the slow logs I got is caused by the time-cost after extra_info.

      Attachments

        1. diagnostic.data.7z
          96.17 MB
        2. mongod.log.2017-04-10T02-19-56.zip
          1024 kB

        Activity

          People

            mark.agarunov Mark Agarunov
            wolf_kdy deyukong
            Votes:
            0 Vote for this issue
            Watchers:
            9 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: