Need information regarding DevTest 9.1 High Availability (HA) mode Installation

[kapoor.s]

Hi Guys,

 

I've few queries regarding CA DevTest Installation and it would be great if someone can help to address those.

 

Query 1: Is it possible to install DevTest 9.1 (SV, App Test) in High Availability (HA) Mode using Load balancers? If yes, can someone please help me with installation guide/details?

 

Query 2: If HA mode isn’t supported by DevTest, then what is the best possible way to provide teams highly available CA DevTest Infrastructure (mainly SV, App Test) so that they can seamlessly use this product, without imposing undue load?

 

Note: I want to setup a DevTest Lab and have 2 servers with me.

 

Thanks,

Siddharth

[Anon Anon]

Let me try:

There is no OOTB HA mode in SV, having said that you can leverage one of your fav loadbalancers and route traffic to different VSE's end points ( ex a virtualized REST or WS endpoint) . It is always a good practice to have multiple VSE instances running to make sure they are ready to serve on demand. Think SV (VSE server) as any of your typical application service providers without all the slack and maintainence that comes with it, you may also want to create a ping service ( ex a dummy rest service that accepts a get call) or leverage invoke API to query vse health check and if down take appropriate actions, a little bit of simple and creative solution goes a long way in making sure your fav virtual services are available to all of your customers ( internal/external)

[kapoor.s]

Hi Bhavin, thanks for your reply! I've few queries based on your response, which are as follows:

 

1. Have you ever setup any load-balancer at your end to route traffic to diff VSE's? If yes, which one? Sorry, but load-balancers are new to me.

2. If I want to use a load-balancer, does it require any configuration/setup at DevTest end?

3. Say I have 2 VSE's (both on different servers) and are connected to a single registry. Now if I use a load-balancer, do I need to have 1 VSE up and other down, which is as per active-passive mode? Or can I have both VSE's up and running (with VS deployed on both), and load-balancer can route traffic between these 2? Just want to know how does this logic work?

 

Really appreciate your help!

[Koustubh Warty]

Just as an FYI....I have used the AWS ELB to send requests to multiple VSEs. ELB is a software load balancer for AWS. All the configuration needs to be done in the ELB itself where you specify the details on the VSEs. Then ELB decides which VSE to route the request to.

[kapoor.s]

Thank you!

[Anon Anon]

1 - Have you ever setup any load-balancer at your end to route traffic to diff VSE's? If yes, which one? Sorry, but load-balancers are new to me.

No, but I have worked with customer SME who have set LB in front of VSE's, but there's always first time for anything

https://tomcat.apache.org/tomcat-5.5-doc/balancer-howto.html

 

2 - If I want to use a load-balancer, does it require any configuration/setup at DevTest end?

None that I am aware of, VSE's are indifferent to LB's

 

3. Say I have 2 VSE's (both on different servers) and are connected to a single registry. Now if I use a load-balancer, do I need to have 1 VSE up and other down, which is as per active-passive mode? Or can I have both VSE's up and running (with VS deployed on both), and load-balancer can route traffic between these 2? Just want to know how does this logic work?

If I were you, I would get hold of my LB SME, ask him how LB's are configured for existing servers and use that strategy for VSE's or atleast you'll have some "design/deployment" idea's

 

I wish I had some better answers for you

 

 

 

 

 

[kapoor.s]

Thanks Bhavin!

[Anon Anon]

Thanks Bhavin. Me and Siddharth are part of same project. I have few more questions on HA implementation. Here our customer requirement is bit different. Current deployment architecture is -

---------------------------------------------------------------------------

Server 1 : Enterprise Dashboard, VSE

Server 2 : Registry, VSE, Portal, Coordinator, Simulator

---------------------------------------------------------------------------

Server 3 : Enterprise Dashboard, VSE

Server 4 : Registry, VSE, Portal, Coordinator, Simulator

---------------------------------------------------------------------------

The environment we want enable HA will be accessed by developer and tester from feature teams and they will also have a deployment control. As per your response if we load balance VSE's endpoint then how come developer will know which VSE to use for service deployment from Workstation ? Are you suggesting to provide loadbalancer URL in the DevTest configuration and control the deployment ? Any suggestion on deployment architecture ? Your help would be appreciated. Thanks.

[Kenneth_Avery]

There may be better options, but I would likely give my VSE a name that would be meaningful to the users so that they could recognize which VSEs should be a part of their deployment (VSE1PerfEnv1) or whatever works for you.

[Anon Anon]

Couple of things:

Dev/Design time

1 - Make sure you set appropriate rights/access controls on who gets to deploy on what VSE's - you can use this -  Resource Groups - DevTest Solutions - 9.1 - CA Technologies Documentation  feature so that you can control VSE access

2 - Name your VSE's appropriately

 

RunTime (SV consumption)

1 - End users/SV consumers access SV endpoints thru LB's based on how the LB's are configured/setup