Resource requirements for containerized masking

Virtual CPUs

Container Name

Requirements

Notes

app

8 vCPUs

  • CPU resource shortfalls can occur both on an over-committed host as well as competition for host resources during high IO utilization.

  • CPU reservations are strongly recommended for the Containerized Masking so that Delphix is guaranteed the full complement of vCPUs even when resources are overcommitted.

mds

4 vCPUs

proxy

1 vCPUs

Memory

Container Name

Requirements

Notes

app

128 GB or higher vRAM (recommended)

  • 64GB vRAM (minimum)

  • The masking service on the Containerized Masking Engine uses its memory to process database and file blocks.

  • More memory can sometimes improve performance. Memory reservation is a requirement for the Delphix VM.

  • Overcommitting memory resources will significantly impact the performance of the Containerized Masking.

  • Reservation ensures that the Containerized Masking will not stall.

mds

32 GB or higher vRAM (recommended)

  • 16GB vRAM (minimum)

proxy

1 GB vRAM

Metadata storage

The minimum recommended storage size is 50 GB.

Application Settings for containerized masking

  • MaximumMemoryForJobs

    • Setting the property to MaximumMemoryForJobs = App container max memory limit - 4GB is recommended.

      • If the property MaximumMemoryForJobs = 0, then the system will calculate available memory automatically using this formula: 75% of the App container max memory limit - 6 GB

  • NumSimulJobsAllowed

    • The default value is 7.

    • Recommended to set this to 0 (the max number of jobs that can be scheduled in parallel is max running jobs = App container’s total available core - 1).

For further details on updating this setting, refer to Application settings APIs page.

Important notes

  1. If a limit in pod resources is not defined, the system will not be able to calculate MaximumMemoryForJobs correctly. In that case, you would need to set the value for that property manually to avoid memory issues.

  2. Resource utilization will be affected by job configurations, like min/max memory, stream, and number of columns/tables to mask, etc. Therefore, a job’s min/max memory, streams, and App Container’s memory limit must be adjusted accordingly.