We have a strange issue on our Viprion environment.
For the CVE vulnerability we had to upgrade our software version from BIGIP-18.104.22.168-0.0.4 to BIGIP-22.214.171.124-0.0.5.
The standby hypervisor and guests was upgraded and fail over was completed.
However we found that on one of the guests one of the applications failed. (Other applications on the guest works)
It seems that the floating IP is not used in communication for the application traffic.
All other guests (five guests) on the hypervisor works without any problems
Once the software is failed back to the BIGIP-126.96.36.199-0.0.4 the application starts working when the same process is followed.
All behavior seemed to be pointing to a software bug however we performed another test by upgrading to version 15.
After the upgrade we still experienced the same issue when the standby becomes the active guest
Floating is set to traffic-group 1 on both
Configs are in sync
VLAN's are the same
If there was a configuration difference the application should also fail when running on the BIGIP-188.8.131.52-0.0.4 version
We do use route domains and there is an irule on the configuration
Because the application is configured in a partition for its route domain I do not think we need to specify a % if we single out a node in the irule
Any suggestions or has someone run into the same problem?
This seems probably too obvious to be the cause but, just in case, have you checked the virtual address (Local Traffic > Virtual Servers > Virtual Address List) for the application that fails to make sure that it is in traffic-group-1 rather than traffic-group-local-only?
It is not part of the local traffic-group so it should float across when fail-over occurs.
As I mentioned if they are running on version BIGIP-184.108.40.206-0.0..4 the fail-over works perfectly