0

I have been preparing for a TrueSight deployment for some time now but would like to share my high level architecture with everyone and get some feedback as well as some questions I would like to solidify before continuing.

The requirements in our environment are:

  • Service Provider to Tenant Configuration (We are a managed services provider and will eventually monitor customer environments but want to start with our own environment first)
  • Highly Available Components
  • Disaster Recovery

Modules:

Atrium SSO
Integration Service Host
Presentation Server
Infrastructure Management Server

Servers:

Atrium SSO Server (2)
Integration Service Host Server
Load Balancing Server
Infrastructure Management Server (2)
Presentation Server (2)

I deployed 8 VM's to prepare for the roles listed above in a dev environment and installed Linux (Cent-OS) on all the machines. I am currently going through the scattered documentation trying to build an installation procedure and making sense of everything - Here is where I can use some help.

Here is what I have so far (feel free to correct me or provide suggestions):

1. Configure Load balancing server

  • As far as I can tell there is no guide to do this, only high-level end-requirements as follows:
    • A third-party load balancer with reverse proxy configuration support
    • Configure the load balancer such that the active node is determined based on the HTTP response of 200 and all data is routed to it.
    • Ensure that HTTP session stick mode (sticky bit) is ON
    • Ensure that load balancer supports HTTP & HTTPS communication
  • I downloaded HAProxy on the load balancer VM but have not yet configured.
  • Question #1: Is there a way to make the load balancer highly available? Also DR?
  • Question #2: Is there documentation outlining how to do this (other than the very general HAProxy documentation).
  • Question #3: Can one load balancer be used between the 2 Atrium SSO VM's as well as the 2 infrastructure management servers, or do I need to setup a second load balancer for those two?

2. Install Atrium SSO

  • First set up a shared storage cluster between the two SSO Linux VM's
  • Question #4: Is there any guide on how to do this?
  • To perform a HA installation follow the directions listed in the attachments here: https://docs.bmc.com/docs/display/public/sso90/Installing+BMC+Atrium+Single+Sign-On+as+a+High+Availability+cluster
  • Verify installation by following steps also outlined in link above
  • Perform post-installation tasks and configuration as outlined here: h*ttps://docs.bmc.com/docs/display/public/sso90/Configuring+after+installation
  • Question #5: Are these documents up to date? Last update was 2015...
  • Question #6: Are there any DR instructions for Atrium? Could not find anything mentioning DR...

3. Install the Presentation Server

  • Follow pre-installation instructions and repeat on 2nd VM here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Preparing+for+the+Presentation+Server+installation
  • Follow installation instructions here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Performing+the+Presentation+Server+installation
  • Follow HA post-install configuration instructions on the load balancer: h*ttps://docs.bmc.com/docs/display/TSOMD107/Configuring+aliases+and+hosts+for+High+Availability+mode Note: Recent comment on this article seems to indicate the commands provided may not work
  • Verify the Presentation Server installation and operation by following instructions here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Verifying+the+Presentation+Server+installation+and+operation
  • Register Atrium SSO w/ Presentation Server by following instructions here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Registering+Atrium+Single+Sign-On+with+the+Presentation+Server
  • Configure Disaster Recovery by following instructions here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Configuring+a+disaster-recovery+system+for+Presentation+Server

4. Install the Infrastructure Server

  • First set up a shared storage cluster between the two Infrastructure Server Linux VM's, back to Question #4: Is there any guide on how to do this?
  • After setting up the cluster, perform the following configuration? Not sure about this: h*ttps://docs.bmc.com/docs/display/TSOMD107/Setting+up+the+cluster+in+Linux+7.x
  • Make sure all pre-installation tasks have been completed as outlined here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Preparing+to+install+Infrastructure+Management+in+high+availability+cluster+mode
  • Follow the installation instructions for installing in HA cluster mode here: https://docs.bmc.com/docs/display/TSOMD107/Installing+the+Infrastructure+Management+Server+in+HA+cluster+mode+on+Linux
  • There is conflicting information on HA for this module. As of March 23rd, 2017 the following info has been added: "HA is supported only with an Oracle database. You cannot configure HA with the embedded SAP SQL Anywhere database."
  • Question #7: I don't believe this was the case in the 10.5 documentation, has this changed? If so, why? Does "not supported" mean it is not possible to achieve or just not officially supported anymore by BMC?
  • Enable Disaster Recovery by following instructions here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Installing+Infrastructure+Management+in+the+disaster+recovery+mode

5. Final Touches

  • Register the component products (e,g, Infrastructure Management Server) with the Presentation Server by following this article: h*ttps://docs.bmc.com/docs/display/TSOMD107/Registering+the+component+products+with+the+Presentation+Server
  • Further integrate with other components (e.,g. Atrium SSO) by following articles listed here: h*ttps://docs.bmc.com/docs/display/TSOMD107/Integrating
0

Why are you deploying such a complex system to do infrastructure monitoring? There are a lot of great far more simple tools out there for a lot less money. Almost anyone will tell you to go low cost on the infrastructure side (the strength... maybe) of BMC, and go heavier on the APM side of things. Most of the tools in this suite are extremely old, and very dated. You'll find administrating them a major challenge.

  • It's not just for internal infrastructure monitoring. I am setting it up as a Service Provider to tenant deployment, meaning we will be monitoring tons of customers using this deployment (hence all the HA and DR complexity). As far as I am aware, the tools are all updated and the latest versions are not more than a couple months old. Regardless, the software has been purchased by employing company and they need it deployed. – SirLearnAlot May 16 '17 at 19:13
  • Good luck. The server monitoring product (Patrol) was created in the early 1990s, making it over 20 years old. You'll spend a ton of time trying to get those tools working. Usability is a whole other issue. Anyways good luck with the deployment, I haven't done it in many years since open source tools have evolved considerably. – Jonah Kowall May 18 '17 at 13:58
  • ... I don't understand your logic. Windows was released back in 1985 it doesn't mean that new versions have not been released since?! Can you compare Windows 1.0 to windows 10? Similarly Patrol has been updated to multiple new releases since the 1990's.... – SirLearnAlot May 19 '17 at 14:06
  • The architecture remains the same, Windows has had many major rewrites, and new kernels/UIs applied. This is not the case with the BMC tooling. They do not invest in technology in the same manner. Their owners (private equity) limits investment in products in order to return cash to investors. It's time to wake up and understand how enterprise software works, – Jonah Kowall May 20 '17 at 10:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.