Test Data Manager

Expand all | Collapse all

Number of Repository with Portal

Jump to Best Answer
  • 1.  Number of Repository with Portal

    Posted 10-01-2018 02:29 PM

    Hello,

    Is it possible to have more than 1 instance of the portal install on the repository?

    In our system we would like more than 1 instance of the portal (different server) but only to 1 GTREP database?

    Thanks,

    Francois



  • 2.  Re: Number of Repository with Portal
    Best Answer

    Posted 10-01-2018 04:19 PM

    This has never been officially tested. Somethings may not work properly or at all.

     

    For example, we store the URL to portal in the GTREP. If you have multiple instances, then selecting from this table will potentially return too many results and cause issues with the components that are attempting to get the portal URL.

     

    For example:

    When datamaker attempt to run an enterprise publish, in a best case scenario all jobs will be sent to the first portal server.

     

    I would recommend opening an idea with some additional details regarding how you would like to utilize 2 portal instances on a single repo.



  • 3.  Re: Number of Repository with Portal

    Posted 10-03-2018 08:58 AM

    Thanks for the feedback.



  • 4.  Re: Number of Repository with Portal

    Posted 10-04-2018 04:14 AM

    Francois

     

    Out of interest - what is the requirement behind the request for multiple instances of the Portal on the same repository?

     

    • Is it scalability / fault tolerance (HA / DR) / some other reason … ?

     

    It's possible that there may be other approaches that might achieve what you need.

     

    Regards,

                  Adrian



  • 5.  Re: Number of Repository with Portal

    Posted 10-04-2018 02:22 PM

    Hello Adrian,

     

    Our main reason is scaleability, having 1 set of codes that you can call from multiple servers. You can create one set of connection, you don't have to replicate them on all our servers. We also wanted all of the log to be at one place. No need to search on what server was the job run. We have to provide reports that provide the what, when, where of masked data. 

     

    For sure DR, if the server is down we can have another one take the lead.

     

    Thanks

     

    Francois