HCL Domino Ideas Portal

Welcome to the #dominoforever Product Ideas Forum! The place where you can submit product ideas and enhancement request. We encourage you to participate by voting on, commenting on, and creating new ideas. All new ideas will be evaluated by HCL Product Management & Engineering teams, and the next steps will be communicated. While not all submitted ideas will be executed upon, community feedback will play a key role in influencing which ideas are and when they will be implemented.

For more information and upcoming events around #dominoforever, please visit our Destination Domino Page

Add a new idea Subscribe
Status Needs Clarification
Workspace Domino
Categories Administration
Created by Guest
Created on Jul 24, 2024

RELATED IDEAS

Designer Task performance improvements

We are facing situation when design task on 1600 custom dbs ran for more than 8 hours while doing full design upgrade. After checking different settings, it turned out that in runs 10x faster when the database is not transactionally logged. HCL Support suggested that it's working as designed, but to raise this as an enhancement request (case CS0639072).

As a workaround, we are now running the design update with indirect files, running 5 design tasks in parallel, cutting down the time to less than 2 hours, but it requires creation and maintenance of the ind files.

Is it possible to improve the standard design task performance by:

  • Improving the performance on transactionally logged dbs?

  • Having standard parameter to run the design task with multiple threads (each handling different dbs)?





  • Attach files
  • Admin
    Thomas Hampel
    Reply
    |     Aug 22, 2024
    Could this be caused by a bottleneck in your disk i/o? Certainly translog causes more transactions overall but the limiting factor might be disk performance?
    1 reply
  • Guest
    Reply
    |     Jul 27, 2024

    OK tested on my own and for me it is 2,5 times slower with transaction logging enabled.
    I tried my own design task multi threaded. It should not make a difference if translog is enabled. I also tried the standard design task.

    [ Daniel Nashed ]

  • Guest
    Reply
    |     Jul 25, 2024

    Hi
    I don't think translog should make any difference to the performance of the design task.

    If you do the math it takes around 18 seconds per database. This is what I also got in my tests with a multi threaded server task on a standard Linux machine refreshing the mail database template after having it signed.

    Yes it would make a lot of sense to run the signing in parallel. And it would also make sense to redirect the log output and not have all of it in the log of the server.


    The DesignRefresh API call has a signal proc. but sadly that has no call-back parameter.
    So when operating multi-threaded this could become tricky to implement with logging.

    For now I am logging to the console in my test via printf. This avoids the logging overhead.

    Yes I think the idea to have it multi threaded makes a lot of sense.
    But I don't think translog has any impact here.

    You are actually saying it would have taken 2 seconds for a larger application to have the design refreshed if I get that right?

    How was this tested?


    [ Daniel Nashed / https://blog.nashcom.de ]


Copyright © 2023 HCL Technologies Limited