Addressing the single point of failure

[Mecanik]

Hi,

I've been in touch with support for quite some time now, because I'm planning on using ADC for a HA setup. However, there is a dilemma unanswered; Load balancers should not have a single point of failure. Running ADC without replication does not address this, so what is the explanation?

Please provide me some valid points as to why ADC can be a "good" load balancer (besides the awesome features), and how can it even be a load balancer in the first place. Suppose you have ADC and 5 LSWS, if the ADC stops for whatever reason it's game over.

Thanks

 

[serpent_driver]

LiteSpeed ADC is a load balancer, but isn't a self watching application that checks itself if it is running. I think you have a wrong understanding about what a load balancer is good for.

 

[Mecanik]

LiteSpeed ADC is a load balancer, but isn't a self watching application that checks itself if it is running. I think you have a wrong understanding about what a load balancer is good for.

I think you do not understand what a Load Balancer is tasked with, what it does, what it's supposed to do, and why it should be highly available. My question wasn't about "self watching application" (what is that even), but about the design architecture behind ADC.

Single point of failure in a load balanced environment makes absolutely no sense to me, as if it goes down if any reason, the whole infrastructure is dead. Which is why most hosting companies provide Load Balancing services that are replicated and come with SLA 99.95% at least, using HAProxy and other open source software.

The only valid and current usage for ADC might be (correct me if I am wrong, still waiting for staff to reply), is Fail-Over. If back-end A is dead, steer traffic to back-end B, and so on. Why? Because there are many areas that are not covered to use ADC as true "load balancing".

To summarize, I am waiting for staff to reply with some valid points as how can one use ADC in a large scale deployment to achieve High Availability.

PS: Stop being toxic in every post you make, you seem like a troll.

 

[serpent_driver]

The only one talking about toxic things here is you and as long as you are hostile to other users with this choice of words, you won't get any help. I will therefore repeat what I have already written. LS ADC is a load balancer, but a load balancer does not check itself to see if it is still alive. If you want to check the availability of the server, then you need another and perhaps external solution that monitors the status of the server. But you can also set up your bed next to the ADC server and monitor this server.

Please stop treating other users unkindly.