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

Failover for specific virtual server

VictorRuiz
Altostratus
Altostratus

Hello everyone,

I'm kinda new on F5 and i'm trying to learn about different failover mechanism.

Let's supose we have an active/standby configuration and each device have 3 virtual server ( A, B and C )each one with their respective pools.

It is possible to trigger a failover event only if virtual server A goes down? i dont care if B or C goes down.

I'm aware of the HA groups and the possibility of trigger a failover event if some pool members are down but i want to know if it is possible for trigger a failover for a specific pool.

 

 

1 ACCEPTED SOLUTION

Hi @VictorRuiz , 
you can do it by HA score configuration , Configure it for the pool that assigned to required virtual server. 
I see there is no difference to failover based on virtual server status (UP/Down) because if the virtual server’s pool became down , all traffic destinated to this virtual server would be resetted and F5 marks the virtual server down. 
so I see you can confugure HA Group Score. 
to configure it follow these article : 
https://support.f5.com/csp/article/K42542000
in this article you will find the videos that @mihaic sent to you , and I think that this is the only way to meet your requirement. 
I want to add. 
> you can Isolate your configuration by creating 2 traffic groups , one for virtual server B and C whereas the other for your targgeted virtual sever A , let we name it (Virtual_server_A traffic group) . 
> you will create new 2 floating ips address for traffic group 2 and put both of  them in Virtual_server_A traffic group , both of floating ips from external and internal vlan subnet respectively ( i mean client/server sides) to  listen and forward traffic for the Virtual_server_A traffic group. 
also go to (local traffic >> virtual address >> and put your targgeted virtual server A on Virtual_server_A traffic group.
> you can make this traffic group (Virtual_server_A traffic group) Active on the main F5 node with the other traffic group ( virtual server B , C ) , I mean both of traffic groups will be active on one node. 

> until here , I did not make any differnece only created  new traffic group "Virtual_server_A traffic group" and two floating ips (one for client side from external vlan subnet and the other from server side from internal vlan subnet). 

> After that  follow @mihaic  video or article K42542000 , as you will  go to newly created traffic group (Virtual_server_A traffic group) and configure change its options to failover based on HA Score and create an new HA group score based on Pool and give it a score. 

> the effect of this , you will isolate your issue and get a more granular control on "virtual server A" far from the rest of other virtual servers. 

Regards. 

_______________________
Regards
Mohamed Kansoh

View solution in original post

3 REPLIES 3

mihaic
MVP
MVP

here are some details:https://support.f5.com/csp/article/K16947

But: "A pool object should not be the only element used to calculate the health score"

Here is a video on how to configure it: 

https://www.youtube.com/watch?v=hNZmsOU18Sc&list=PLyqga7AXMtPPi-MPCs8eC2b3EZDNqHuBO

 

Hi @VictorRuiz , 
you can do it by HA score configuration , Configure it for the pool that assigned to required virtual server. 
I see there is no difference to failover based on virtual server status (UP/Down) because if the virtual server’s pool became down , all traffic destinated to this virtual server would be resetted and F5 marks the virtual server down. 
so I see you can confugure HA Group Score. 
to configure it follow these article : 
https://support.f5.com/csp/article/K42542000
in this article you will find the videos that @mihaic sent to you , and I think that this is the only way to meet your requirement. 
I want to add. 
> you can Isolate your configuration by creating 2 traffic groups , one for virtual server B and C whereas the other for your targgeted virtual sever A , let we name it (Virtual_server_A traffic group) . 
> you will create new 2 floating ips address for traffic group 2 and put both of  them in Virtual_server_A traffic group , both of floating ips from external and internal vlan subnet respectively ( i mean client/server sides) to  listen and forward traffic for the Virtual_server_A traffic group. 
also go to (local traffic >> virtual address >> and put your targgeted virtual server A on Virtual_server_A traffic group.
> you can make this traffic group (Virtual_server_A traffic group) Active on the main F5 node with the other traffic group ( virtual server B , C ) , I mean both of traffic groups will be active on one node. 

> until here , I did not make any differnece only created  new traffic group "Virtual_server_A traffic group" and two floating ips (one for client side from external vlan subnet and the other from server side from internal vlan subnet). 

> After that  follow @mihaic  video or article K42542000 , as you will  go to newly created traffic group (Virtual_server_A traffic group) and configure change its options to failover based on HA Score and create an new HA group score based on Pool and give it a score. 

> the effect of this , you will isolate your issue and get a more granular control on "virtual server A" far from the rest of other virtual servers. 

Regards. 

_______________________
Regards
Mohamed Kansoh

VictorRuiz
Altostratus
Altostratus

Thanks to both for the answers, i will try the solution on my lab enviroment.

Regards.