Cloud Pak for Business Automation

 View Only
Expand all | Collapse all

Performance impact in CP4BA with Open shift

  • 1.  Performance impact in CP4BA with Open shift

    Posted Tue July 05, 2022 07:33 PM
    Hi All,
    We have CP4BA environment with Open shift. We are using only CPE and ICN. We have 3 Master nodes and 6 Worker nodes (16 vCPU and 32 GM RAM) for ICN and CPE. 3 CPE pods deployed and it can scale up if needed.

    We are planning to migrate 94 million documents from CMOT to FileNet env. and load 24/7.

    Can we load 2 to 3 million document per day with 6 worker nodes?. Would any one share your suggestions. Thanks in Advance.

    Regards,
    Durga.

    ------------------------------
    Durga Prasad Chinthala
    ------------------------------


  • 2.  RE: Performance impact in CP4BA with Open shift

    IBM Champion
    Posted Wed July 06, 2022 08:06 AM
    Hi Durga,

    So, you should be fine, but you need to consider the sizing per CPE pod, remember that the CRD specifies not only the number of CPE pods (i.e. the number of instances of CPE in the cluster) but also the size of each pod (the number of CPU cores and quantity of RAM allocated to it). I'd recommend starting by ensuring that each CPE has 2-4 cores and 8-16 GB RAM allocated​ per pod. The number of workers isn't as important as the size and number of CPE pods. You should be sure that your workers, though, aren't overallocated from a resource-perspective.

    The other factors to consider are what tool you're using to do the ingestion, if you need to do any in-flight format conversions, or any other transformations that might slow-down the process. Assuming you mean CMOD as the source, file-format conversion may be a real concern. You also need to consider network bandwidth, average document size and performance on the extract side.

    That said, in general, 2-3M docs per day is reasonable. 

    Best,
    Eric

    ------------------------------
    Eric Walk
    Director

    O: 617-453-9983 | NASDAQ: PRFT | Perficient.com
    ------------------------------