Open jedelman8 opened 7 years ago
I would just say to wait till OC will provide the models for action commands and provide a proper implementation. These methods have been designed improperly from the beginning (the same, I have it) and I think we'd only continue on the same line doing so.
Fair enough if that's the consensus...module works and my use-case has been accomplished...so for now, it'll remain a custom module!
@mirceaulinic are there any OC models for actions that you know of? I haven't seen them (haven't looked for them either)
Currently there aren't any models available yet, but there will be something in the public repo soon-ish :-)
Something tells me the model is going to be even messier than the platforms model...
Are there any news regarding this issue?
Currently I get completely different answers from the same unreachable destination using Junos and IOS what makes further processing more complicated. Also is the "probes_sent":0 and "packet_loss": -2 on IOS expected result in case 2 echo request were sent and no reply was received?
Using napalm_ping module, count:2, destination: 4.4.4.4
Result on IOS: ... "item": "4.4.4.4", "results": { "success": { "packet_loss": -2, "probes_sent": 0, "results": [], "rtt_avg": 0.0, "rtt_max": 0.0, "rtt_min": 0.0, "rtt_stddev": 0.0 } }
Resutl on Junos: ... "item": "4.4.4.4", "results": { "error": "Packet loss 100" }
@jemcek would you mind opening an issue on napalm-ios
. Looks like the output returned by IOS is not compliant with the model we specified so we need to fix that. Thanks!
@mirceaulinic do you have any news about the "action" model?
done.
No, I don't have any news yet. I guess is going to take longer than I thought...
Reference: https://github.com/napalm-automation/napalm-ansible/pull/53
one thing I'm thinking about within the ansible module and @dbarrosop sugessted maybe we just update the
ping
method within NAPALM so we can keep the ansible module a wrapper.Normalizing the key used regardless if success or error.
Today it's this:
I'd rather see a status key that has success or error as values. Then a msg key for the error message and results for the dict of results. What are your thoughts?