cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 12

Solarwinds in Azure with HA

We are planning to deploy Solarwinds in Azure along with HA. This is the architecture we have designed. Can some one help me if this is the right one or any changes need to be made.

Modules we have are NPM, SAM, NTA, VNQM, WPM, NCM, IPAM, UDT, Vman & Toolset.

Below are the considerations we took during the design.

  • Entire Solarwinds environment is planned to be in single Vnet.
  • HA functionality need to be enabled in primary polling engine and 3 additional polling engines
  • HA will be enabled using virtual hostname
  • Each polling engine servers and respective HA servers are placed in its on availability set and are in same subnet
  • Orion DB servers and NTA DB servers are in SQL clustered mode

Solarwinds Architecture_Azure.JPG

6 Replies
Level 8

@sreenathmp 

Hi ,

We have SolarWinds setup in Azure platform. could you please provide how can we achieve and Configure SolarWinds application HA in Azure LB and how it will work during failover.

0 Kudos
Level 17

  • Each polling engine servers and respective HA servers are placed in its on availability set and are in same subnet

This statement concerns me, as your HA pairs will be 1 Active and 1 Standby. So polling will not happen from both servers that are paired, just one or the other (I think you already know this detail though).

Load Balancer in front of the web engines can work, but I am curious what the effort is with the LBs to the Databases.

Everything else looks good.

0 Kudos

Hi Cahunt,

Polling engines and respective HA servers will be in active passive mode only as Solarwinds HA functionality is used even though they are in same availability set.

Regarding LB in database, i think for clustering in Azure LB is required(I am not a cloud expert). Please advice if this is correct or not.

0 Kudos

Yes Sir, I did double check and support for the Availability group listener is enabled with the LB in place.

Regarding the servers, there is a 1 to 1 need/ratio (Primary/Secondary) services will be active from one of the two servers within each pair.

You can choose which server should be active if there is a preference, and also a further setting to prevent fail back before a specific time frame.

That option can give your failed server time to restart or self heal and get back to normal before the system even tries to fail back to a set primary server, and prevent fail over flapping.

0 Kudos

Hi CA hunt,

I hope you are explaining about the HA setting given below.

HA.JPG

That is exactly the page I was referring to. You got IT!

0 Kudos