Versions Compared

Old Version 5

changes.mady.by.user Brian Miller

Saved on

compared with

New Version 6

changes.mady.by.user Brian Miller

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Verification: Add Scenario2 checks for Cassandra batch_size thresholds + Akka performance tuning enhancements

...

  1. Please follow the instructions on the Migration page but use SC 3.5 SP1 everywhere where it references SC 3.5.

...

Verification

  1. Scenario 2 ONLY: Confirm batch_size_warn_threshold_in_kb and batch_size_fail_threshold_in_kb in /etc/cassandra/default.conf/cassandra.yaml have been bumped up from their defaults, ex:

    Code Block
    batch_size_warn_threshold_in_kb: 250
batch_size_fail_threshold_in_kb: 300

    If you make changes to the above, restart the Cassandra service:

    Code Block
    sudo systemctl restart cassandra

    (info) If you frequently deal with large artifact sizes you may want to configure your monitoring to watch for the “batch size” warn messages and bump up as needed.

  2. Scenario 2 ONLY: Make sure sc-store service’s application.conf at /opt/icx/syndeia-cloud-current/store-impl-3.5-SP1/conf/application.conf contains the following performance tuning enhancements on ~L95:

    Code Block
    cassandra-query-journal {
  refresh-interval = 25ms
  
  # added events-by-tag nesting during lagom upgrade to 1.6.5 (to cater to akka persistence cassandra migration to 0.80)
  # https://doc.akka.io/docs/akka-persistence-cassandra/0.101/migrations.html
  events-by-tag {
    # Turn off the eventual consistency delay
    eventual-consistency-delay = 50ms

    # Turn on event by persistence id tracking, this timeout is the maximum amount of time that akka-persistence-cassandra will hold off publishing an event when missed events are detected
    # changed name from delayed-event-timeout to gap-timeout during lagom upgrade to 1.6.5 (to cater to akka persistence cassandra migration to 0.80)
    # https://doc.akka.io/docs/akka-persistence-cassandra/0.101/migrations.html
    gap-timeout = 30s 
    flush-interval = 25ms
    first-time-bucket = "20221101T00:00"
  }
}

    If not, add them and restart the sc-store service.

    Code Block
    sudo systemctl restart sc-store

  3. Follow the steps in the section Validating Syndeia Cloud Installation & Configuration on the following page to verify your installation: