On 16.8.2021 19.23, [email protected] wrote:
A simpler might be to mark Duo dead for a configurable number of seconds
after which it's marked as alive again without a check. The next
authentication would then either work or again trigger marking it as dead.
This solution is now in 4.25 patches and available with patch packages.
What now happens is that when CheckTimerInterval is set to 0, API
failure sets the API status to failed as usually. When
FailureBackoffTime, a new parameter, seconds have elapsed since the
failure was detected, the API is considered to be working again.
In other words, this matches the behaviour quoted above with
FailureBackoffTime being the configurable parameter for the number of
seconds.
The default value of FailureBackoffTime is 60. If set to 0, then the API
never remains in the failed state. Longer values avoid using the API,
for example, when there's a need to access many different devices to
troubleshoot timeout causes.
As always, feedback is welcome.
Thanks,
Heikki
--
Heikki Vatiainen
OSC, makers of Radiator
Visit radiatorsoftware.com for Radiator AAA server software
_______________________________________________
radiator mailing list
[email protected]
https://lists.open.com.au/mailman/listinfo/radiator
