Technical Forum
Ask questions. Discover Answers.
cancel
Showing results for 
Search instead for 
Did you mean: 

4 BIG-IP in one cluster?

Shurik
Altostratus
Altostratus

Hello guys,

 

We have two data centers, with L2 link between the DCs. Is it possible to create one cluster (2 BIG-IP LTM in primary DC + 2 BIG-IP LTM in standby DC)?

And in case one of the services/links go down in primary location, the VIP for that specific service will be up in standby DC?

Is it a good practice?

 

Thank you.

1 ACCEPTED SOLUTION

Daniel_Wolf
Nacreous
Nacreous

It is sort of a good practice, but to shift traffic to a backup data center you would also use F5 DNS.

You would have two clusters, one in each datacenter. And F5 DNS would distribute traffic across data centers.

Take a look at this video: BIG-IP DNS Load Balancing Introduction

Otherwise you would have plenty of (virtual) hardware sitting idle.

 

 

View solution in original post

5 REPLIES 5

Daniel_Wolf
Nacreous
Nacreous

It is sort of a good practice, but to shift traffic to a backup data center you would also use F5 DNS.

You would have two clusters, one in each datacenter. And F5 DNS would distribute traffic across data centers.

Take a look at this video: BIG-IP DNS Load Balancing Introduction

Otherwise you would have plenty of (virtual) hardware sitting idle.

 

 

On a second thought - without knowing more about your (cloud) infrastructure, your business requirements, the application itself, ... and so on, it is nearly impossible to provide you with a blueprint of a "silver bullet" architecture that will satisfy all your requirements.

 

From the top of my head, I could see a couple of scenarios.

  1. One LTM cluster of four BIG-IP devices with one or more traffic groups
  2. Two LTM clusters in two datacenters
  3. Four standalone LTM might also be possible, depending on your level of automation / config management

 

Option 2 and 3 combined with F5 DNS or F5 DNS Load Balancer Cloud Service and optional also with BIG-IQ.

 

Depending on how far you have developed this idea already, it might make sense to reach out to someone who can help you with the design of your topology.

Hello Daniel,

 

I think the f5 DNS will be best solution here, but unfortunately it wasn't in the budget... So I'm exploring alternatives designs.

 

Scenario 2:

Is there a way to failover virtual server between two different clusters?

 

Thanks again!

 

 

Hello Shurik.

 

If your idea is that both clusters have the VIP configured and one cluster has the VIP up, while the other cluster has the same VIP down, like a warm standby... nope. At least to my knowledge. Probably possible with some external tools for monitoring & scripting.

With your requirements and budget a 4 node cluster might be a good solution. But also I am not seeing the full picture of your requirements, so I cannot recommend a "best solution".

 

KR

Thank you Daniel! I like the DNS solution, we will see what can be done here 🙂