This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Foglight PI install on windows 2016 server VM best configuration

New to Foglight and the discussion board. Got a newbie question. Looking for insight from someone who has done this before. We have 120 potential agents and so have spec'd the resources needed.

How should it be architected?  3 VMs or one big one?  Broken out like the specs show into 3 servers one for Foglight Manager, one for Agent manager, and one for PI repository?

Or perhaps FMS, Agent Mgr. + PI, FMS repository.

Also any advantages for us to not use the FMS embedded postgres  repository and use sql server instead ?  I can see having external sql server will require another license, patching, upgrading and other maintenance where the Postgres is embedded and taken care of by Foglight.

Thanks

sjb

  • Hello sbenn,

    we have a deployment guide for database agents + PI installation. Look at here: https://support.quest.com/foglight/5.9.3/technical-documents

    Here my tips:

    Have one machine for the FMS, one for the FglAM and one for PI. 

    For PI we recommend Linux as it is a more stable than Windows. If you are using a virtual machine, please make sure that the resources (CPU + memory) are fully reserved for that virtual machine. This is very important as we see PI performance issue again and again if the resource reservation has not been done.

    In the deployment guide, you will find how much memory and CPU you need as this change based on the amount of DBs you want to monitor.

    Regarding the FMS DB - embedded or external:

    I haven't seen any issues with the embedded Postgres DB in the last years. It is stable and performant. You are correct, you do not need an additional license or take care of upgrades and maintenance. But if you have more SQL Server knowledge and you will feel much more comfortable with SQL Server, then you can use a SQL Server. Both DBs are working fine.

    In general, I would recommend assistance from our technical consultants as they can do a proper sizing for you based on your requirements. You can discuss in detail your needs and your plans (how many DBs should be monitored, will this be extended further in the future etc.). To get in contact with the technical consultants, you can contact your Quest account manager and he/she can arrange the contact on your behalf. Or you can use our web portal - look at here: support.quest.com/professional-services-product-select

     Let me know if you have further questions.

    Kind Regards

    Nicola

  • Thanks for the detail. The deployment guide did a good job spec'ing out the VMs but I still wasn't clear on the architecture. So I will need three VMs; one for FMS + Postgres DB , one for FglAM, and one for PI.   If we go SQL Server would it make more sense to have one for FMS, one VM for FglAM and PI, and one for FMS repository (SQL Server) ?  

  • Hello sbenn,

    you can use one server for the FMS and the FMS DB. But this can change depends on the number of agents. Most customers have a separate machine if they use an external DB. 

    But might be better to discuss the best setup with our technical consultants as they have the field experience to give better advice. They can tell much better how much memory + CPU is needed on which machine and how to change the Foglight default settings to the ones you need.

    Kind Regards

    Nicola