Details

    • Type: Technical task
    • Status: Resolved
    • Priority: Major
    • Resolution: Duplicate
    • Affects Version/s: 2.0, 2.0.1, 2.1.0
    • Fix Version/s: 3.0
    • Component/s: ns_server
    • Security Level: Public
    • Labels:
    • Environment:

      Description

      3 nodes cluster/1 bucket/700K/3 ddocs:
      2 ddocs have 2 views each: 1 has good map function another one has infinity loop:
      function (doc, meta) {
      for (var i=0; i != 1.1; i++)

      { emit(meta.id, "infinity"); }

      }

      and 1 ddoc has 1 view with good map function

      Steps to reproduce:
      1. Reboot one node
      2. Fail over it when it's warming up
      3. Add it back
      4 Rebalance

      attaching diags

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

        Activity

        Show
        iryna iryna added a comment - Diags https://s3.amazonaws.com/bugdb/jira/MB-7231/79ffbdda/104-diag.txt.gz https://s3.amazonaws.com/bugdb/jira/MB-7231/79ffbdda/105-diag.txt.gz https://s3.amazonaws.com/bugdb/jira/MB-7231/79ffbdda/110-diag.txt.gz .104 was failed over
        Hide
        alkondratenko Aleksey Kondratenko (Inactive) added a comment -

        This one is interesting. While those 5 seconds of map function timeout are not hit, entire erlang interpreter thread (aka scheduler thread) is busy and given that erlang is unable to properly balance work to other threads it caused timeouts but not due to lack of async IO threads.

        Good test case. Thanks a lot.

        Show
        alkondratenko Aleksey Kondratenko (Inactive) added a comment - This one is interesting. While those 5 seconds of map function timeout are not hit, entire erlang interpreter thread (aka scheduler thread) is busy and given that erlang is unable to properly balance work to other threads it caused timeouts but not due to lack of async IO threads. Good test case. Thanks a lot.
        Hide
        alkondratenko Aleksey Kondratenko (Inactive) added a comment -

        CC-ing Filipe

        Show
        alkondratenko Aleksey Kondratenko (Inactive) added a comment - CC-ing Filipe
        Hide
        alkondratenko Aleksey Kondratenko (Inactive) added a comment -

        And back to me. Looks like a case for VM splitting regardless of async IO threads resolution

        Show
        alkondratenko Aleksey Kondratenko (Inactive) added a comment - And back to me. Looks like a case for VM splitting regardless of async IO threads resolution
        Hide
        dipti Dipti Borkar added a comment -

        This may happen even in case of a reasonable query that loops over an array.

        Show
        dipti Dipti Borkar added a comment - This may happen even in case of a reasonable query that loops over an array.
        Hide
        maria Maria McDuff (Inactive) added a comment -

        aleksey — this may need the file I/O NIF or some other change moving out to 2.1. if you feel u have a fix for it, pls move it back to 2.0.2 for release.

        Show
        maria Maria McDuff (Inactive) added a comment - aleksey — this may need the file I/O NIF or some other change moving out to 2.1. if you feel u have a fix for it, pls move it back to 2.0.2 for release.
        Hide
        alkondratenko Aleksey Kondratenko (Inactive) added a comment -
        Show
        alkondratenko Aleksey Kondratenko (Inactive) added a comment - MB-8668

          People

          • Assignee:
            alkondratenko Aleksey Kondratenko (Inactive)
            Reporter:
            iryna iryna
          • Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Gerrit Reviews

              There are no open Gerrit changes