Uploaded image for project: 'PuppetDB'
  1. PuppetDB
  2. PDB-5203

Script that runs the pdb locust load tests



    • Task
    • Status: Resolved
    • Normal
    • Resolution: Fixed
    • None
    • PDB 7.6.0, PDB 6.19.0
    • PuppetDB
    • None
    • Not Needed
    • Needs Assessment


      The script does:

       - if a commit SHA is given as a parameter, than it checkouts to that and holds the SHA in a variable, needed to name the results folder.
      If no SHA is given, it pulls the latest changes from the 6.x branch and copies in a variable the latest commit SHA.

      • In the locust results folder, create a folder with <current date without time>-<commit SHA>.
      • inside this folder create another two named ingestion-ON and ingestion-OFF
      • inside the two folders create folders 1000, 10.000, 100.000 and 500.000



      • start pdb
      • after pdb started, run locust in the needed results folders (the -csv parameter accepts folder paths) with all queries, with a warmup with one user, another one with -1 user, 10 users and 20 users.
      • go to n1 server and start the benchmark with (30 min run interval, 30% data change) and the appropriate number of hosts, depending on the sandbox that's used.
      • go to the ingestion results folder and start locust again with the same settings.
      • kill the benchmark after locust is done.
      • kill pdb.
      • inside the results folder delete all files except the .html, _stats.csv and _failures.csv (not sure if we need those). Also the warmup files must be deleted.
      • Check the _failures.csv files, and if they are empty, delete them.

      !!!!!!! For the 500.000 sandbox you need to change the main sandboxes path.




            sebastian.miclea Sebastian Miclea
            andrei.filipovici Andrei Filipovici
            0 Vote for this issue
            1 Start watching this issue



              Zendesk Support