Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-6234

We're leaving some disk performance on the table

    Details

    • Type: Bug
    • Status: Resolved
    • Priority: Major
    • Resolution: Incomplete
    • Affects Version/s: 2.0-beta
    • Fix Version/s: bug-backlog
    • Component/s: forestdb
    • Security Level: Public
    • Labels:
      None
    • Triage:
      Untriaged

      Description

      I'm pretty sure this wasn't case few month ago.

      There's around 2x difference in write throughput with barriers on and off.

      That indicates we're doing fsync calls a bit too often.

      My thinking is with right size of transactions we should be able to fsync seldom enough so that effect of fsync (actually waiting until stuff hits platter) is negligible.

      I should also note that this difference can be due to FS metadata ops which we're doing quite a bit as well (all those file size modification). If this difference is due to underlying FS metadata transactions then there's nothing we can do in short term.

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

        Activity

        Hide
        mikew Mike Wiederhold added a comment -

        Duplicate of MB-6232.

        Show
        mikew Mike Wiederhold added a comment - Duplicate of MB-6232 .
        Hide
        alkondratenko Aleksey Kondratenko (Inactive) added a comment -

        That was actually for different and larger problem.

        Show
        alkondratenko Aleksey Kondratenko (Inactive) added a comment - That was actually for different and larger problem.
        Hide
        alkondratenko Aleksey Kondratenko (Inactive) added a comment -

        I think I have to re-open that.

        Show
        alkondratenko Aleksey Kondratenko (Inactive) added a comment - I think I have to re-open that.
        Hide
        drigby Dave Rigby added a comment -

        Assigning this issue over to forestdb (Chiyoung) - the solution for any disk-performance issues is forestdb, and so ep-engine will not be directly addressing this.

        Chiyoung - please feel to close, I just wanted to direct this to the correct component.

        Show
        drigby Dave Rigby added a comment - Assigning this issue over to forestdb (Chiyoung) - the solution for any disk-performance issues is forestdb, and so ep-engine will not be directly addressing this. Chiyoung - please feel to close, I just wanted to direct this to the correct component.
        Hide
        chiyoung Chiyoung Seo added a comment -

        We will address this frequent fsync call issue (i.e., one fsync call per vbucket file) as part of ForestDB integration with ep-engine that will group a set of vbuckets into a single shard file.

        Show
        chiyoung Chiyoung Seo added a comment - We will address this frequent fsync call issue (i.e., one fsync call per vbucket file) as part of ForestDB integration with ep-engine that will group a set of vbuckets into a single shard file.

          People

          • Assignee:
            chiyoung Chiyoung Seo
            Reporter:
            alkondratenko Aleksey Kondratenko (Inactive)
          • Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Gerrit Reviews

              There are no open Gerrit changes