we have 3 datacenter, located on 3 physical site.
each site has his own ip subnet .
2 haproxy server have been installed with keepalived with one virtual ip address on each site. So we have 6 haproxy with exactly the same haproxy.cfg but on three differents site.
we also had 3 dns server , one per site.
we have round robin in order to reach the three virtual ip with three same A record.
the issue is when on site is down, dns continue to allocate the dead virtual ip …
Do you have a tips for this issue ?
this is for lan access.
for external acces to our three virtual ip we have our firewall that can do loadbalanxer for https flow and so testing health check of the three haproxy behind virtual ip.
How do you solve this issue on your firewall? How is the firewall reachable with all those constraints, in your particular failure scenarios? And what is this failure scenario anyway (I assume the datacenter has redundant connections - connectivity and power).
There is no easy fix for this. You need to go back to the drawing board and find out why you are in this situation in the first place.
Unless you have health checking DNS servers, or external health checker that can modify DNS records, both of which are probably workarounds, you are probably not going to fix this.
This is not something that haproxy can solve.
ibm global ip.
ip floating between sites.
So why don’t you do the same for haproxy?