Forum Discussion
NimbostratusStandby F5 runs external health monitor
I have an external health monitor that I have written to verify a webserver is displaying the correct information and down the node if not. When looking over the http logs of the webserver I noticed that the two standby F5 we have in the F5 cluster are also running the external monitor and hitting the webserver as well. Instead of 1 page view per server every X seconds, it is seeing 3. Is this expected behavior and is there a way to disable the standbys from running this external monitor? I would think it wouldn't need to run it until a failover to active.
Hello, this behavior is as expected.
One reason this is beneficial comes down to troubleshooting - if more than one F5 appliances report that the pool member went "DOWN", it's a better evidence to rule out any possible F5 appliance issues. If you want, you can disable the health-check on standby F5s by writing a conditional check in the external script which verifies the appliance active-standby status prior to sending the health-check request.
If the volume of health-check requests is a problem, perhaps it will help you if you increase the health-check interval a little bit? As far as I'm aware, no easy solutions exist to stop health-checks on the standby appliances. Could be wrong in regards to v11.6 or 12.0 releases.
Hannes_Rapp_162Nacreous
Hello, this behavior is as expected.
One reason this is beneficial comes down to troubleshooting - if more than one F5 appliances report that the pool member went "DOWN", it's a better evidence to rule out any possible F5 appliance issues. If you want, you can disable the health-check on standby F5s by writing a conditional check in the external script which verifies the appliance active-standby status prior to sending the health-check request.
If the volume of health-check requests is a problem, perhaps it will help you if you increase the health-check interval a little bit? As far as I'm aware, no easy solutions exist to stop health-checks on the standby appliances. Could be wrong in regards to v11.6 or 12.0 releases.
gsharriAltostratus
Yes as Hannes says this is expected behavior. If the standby unit is not monitoring the nodes/members and some of them are down, then if called upon to go active it will send traffic to an unavailable server until it discovers that it is down. Some clients will experience failed connections for the duration of the monitor timeout.- P_K
Does anyone found how this situation is dealt on 12.1.1?
Hannes_RappHello, this behavior is as expected.
One reason this is beneficial comes down to troubleshooting - if more than one F5 appliances report that the pool member went "DOWN", it's a better evidence to rule out any possible F5 appliance issues. If you want, you can disable the health-check on standby F5s by writing a conditional check in the external script which verifies the appliance active-standby status prior to sending the health-check request.
If the volume of health-check requests is a problem, perhaps it will help you if you increase the health-check interval a little bit? As far as I'm aware, no easy solutions exist to stop health-checks on the standby appliances. Could be wrong in regards to v11.6 or 12.0 releases.
- gsharriYes as Hannes says this is expected behavior. If the standby unit is not monitoring the nodes/members and some of them are down, then if called upon to go active it will send traffic to an unavailable server until it discovers that it is down. Some clients will experience failed connections for the duration of the monitor timeout.
- P_K
Does anyone found how this situation is dealt on 12.1.1?
