Closed blefeuvr closed 6 years ago
I am not sure this is a bug. I think this would be the expected behavior. Especially for not being able to connect, because you could eventually be able to connect. It could be just that the device is down and rebooting, etc.
And if you pip install napalm when it wasn't installed, then it would try to connect, without having to restart the system.
There is also a scenario where the proxy minion starts up, and the pillars haven't been rendered and it takes a try or two before the proxy minion gets the information to then connect.
So in my opinion, this is by design.
@cro @mirceaulinic opinions?
Thanks, Daniel
I agree, I think it's not a bug. It is ugly, granted, but these devices can be recalcitrant and the safest thing to do to try to get a reconnect is what we are doing right now.
OK thanks a lot for the answer, makes things clearer for me, at least I know this is the way it is normally going. I close, thanks again for your awesome reactivity.
We have an issue where we have a significant number of devices down at at one time (these are devices that are used ad-hoc and when not required are turned off), this causes our Salt master to be overloaded as the proxy process restarts.. Any ideas to resolve this scenario?
Description of Issue/Question
I hesitated a lot between issuing this here or in napalm-salt github. But I will start this here. Setup is quiet simple I have a proxy pillar cisco.sls:
A pillar top file :
And then I try to launch the proxy minion cisco1 process from a minion :
And here comes the issue, if there is a problem somewhere, like:
The process enter in an infinite loop trying to refresh grains, failing because device is not connected, and just trying everything anew.
Part of debug:
Versions Report
minion :
master :