UC Reporter install to monitoring node?

[Jonathan_Beaver]

We would like to install all of the Prognosis components (monitoring, management and reporter) then place the SQL Database on the same server. I realize the server will have to have a pretty large footprint from a CPU and Memory and Disk perspective. However, do you see any issues with this configuration? Do you have any customers that have done this?

Regards Jon Beaver

[Adam_Annen]

Hi Jon,

Thank you for posting on the forum. In regards to your question I think it best to answer it in the form of pros & cons.

Pros
Improved Summary completion times
WAN connection issues cannot prevent summaries
Easier to manage 1 server

Cons
If the UC Reporter processes get busy enough and max out the CPU usage across all cores to 100%, it can affect the data coming in from live collectors and cause gaps.
- Both Word and SQL have been seen to be hungry processes

Summary: Considering your situation and the importance of this data, if you give it enough resource it should be fine. Ideally though, SQL would be best separate to a 2nd server on the Customer side as well (I.e. Not across the WAN). I imagine you will only need to do this for 1 or 2 of your customer’s, so it should remain manageable.
Be sure to add this new SQL server to your UC Reporter Database Surveillance solution too (There is no harm this solution monitoring this SQL server across the WAN).

We have had a number of customer's install UC Reporter to their monitoring nodes in the past and it has been a successful environment. Their size however was faily small (1-5 clusters)

[Jonathan_Beaver]

Adam,

I do have a couple follow-up questions.

1. From a server resource perspective, do you have any sizing documentation that could get me close to what I would need for this architecture?
2. You mention installing UC Reporter on the monitoring node. Using this architecture would we no longer have the need for a Management server?

Thank you sir,

Jon

[Adam_Annen]

Hi Jon,

In response to your questions:

1. I would advise to double the current CPU Cores and Memory on that VMWare server.
I.e.
4 Core + 4gb memory becomes 8 Core + 8GB memory
or 8 Core + 8GB memory becomes 16 Core and 16gb Memory.

I would definitely try erroring on the side of generosity in these figures and monitor the CPU and memory afterwards and if it is not utilizing 50% of the values, descrease them slightly.

2. That is correct. Infact, because your data flow upward does go to a Manager with a shared UC Reporter above that, you should removed the network configuration line pushing it up there. ONce we stabnilize the other components and get a top level managing node above your UC Reporter, you could potentially manage this node from there so it does not flow through other UC Reporters.

Let me know if this answers your questions. Feel free to post more questions here on this forum topic.

Regards
Adam

[Jonathan_Beaver]

Ok, one last question. I'm assuming we can take an existing Prognosis monitoring server, rerun the installer to add the necessary UC Reporting components. Or, would we have to execute a net-new install for each customer we want to re-architect?

Thank you,

Jon

[Adam_Annen]

Hi Jon,

There are 2 options here:

1. If you are on a lower version on that server, you can select extra components during the upgrade process (When the component screen is shown).
2. If it is already on the latest version, Run a backup, uninstall, reinstall (And patch), and then restore the backup
- As an extra safety net, if possible, manually backup the whole Prognosis directory as well

Please let me know which one you plan to use here and if you need extra information on the options.

Regards
Adam