Forum Discussion
Hi Pete,
Thank you very much for your reply! So this is what I have done so far.
Option A: WMI: Installed on all servers Load Balancing Method: Dynamic Ratio (node) PG Activation: Disabled
In this setup, All 4 servers start receiving jobs simultaneously and F5 makes sure that the jobs be distributed evenly and the ratio adjusted automatically once one of the servers hits the threshold value for CPU load. This is the functionality I want except for the fact that I want 2 out of the 4 servers to be in standby till the threshold value is reached.
So I tried a different setup.
Option B:
WMI: Installed on all servers Load Balancing Method: Dynamic Ratio (node) PG Activation: Enabled and activated for "Less than 2 members"
Servers A and B have priroity group 2, and Servers C and D have priority group 1. This seems to work ok, except that the PG activation does not happen till one of the serevers A or B is completely unreachable. The health monitor setup for CPU load threshold does not seem to have any impact on PG activation in this setup.
Is it possible to have PG activation criteria to look at the health monitor and the threshold CPU load value, instead of the system default definitions?
Can you please shed some light on how the configuration can be setup from the web server owners' side?
Thanks again in advance ! Sangamesh
- PeteWhiteNov 26, 2013EmployeeThe point that I was making about how the web server owners can manage this is that they can run a script every minute or so that posts to a file which is available on the web server. If all is running well then the output shows UP, if the CPU/memory is above a threshold the output shows DOWN. On the f5 you create an http monitor that retrieves this file and has UP as the receive string and DOWN as the receive-disable string. When it sees DOWN, it will list the server as disabled which means that persistent connections are sent to the server but no new connections. If you want to use the SNMP agent still, maybe you can set the threshold