http://defect.opensolaris.org/bz/show_bug.cgi?id=10860
Summary: system locations should be activated if no other
condition are met
Classification: Development
Product: nwam
Version: unspecified
Platform: ANY/Generic
OS/Version: All
Status: NEW
Severity: normal
Priority: P3
Component: ON daemon
AssignedTo: nwam-dev at opensolaris.org
ReportedBy: Anurag.Maskey at Sun.COM
QAContact: nwam-dev at opensolaris.org
--- Comment #0 from Anurag S. Maskey <Anurag.Maskey at Sun.COM> 2009-08-22
19:33:21 UTC ---
Email thread from John and Alan:
>> Also, I have one question, as marked below for some results. If a location
>> has the condition "ncu foo is-not active", should I expect it to be active
>> if *no* ncus are active? Because in this case, NoNet is being made active.
>> I'm not sure what is correct, but it does seem a bit counter intuitive
>> that a location who's conditions are technically met is not brought online.
>> Allowing that would let the user override no-net behavior.
>>
>
> That's a good question. It brings up the more general question of
> what nwamd should do in the case of "ties": there are two locations
> whose conditions are met, how should it decide which to activate?
>
> I think there is some logic in the condition checking already to
> prioritize matches, so some types of condition matches are "better"
> than others. One of the general principles is that more specific
> matches are better, which is probably why No-Net is being preferred
> in this case: it specifies an address condition, whereas yours
> specifies a more general ncu condition. Alan, please correct me if
> I have this wrong!
>
>
You're dead right - we use heuristics to rate conditions
relative to one another, e.g. an "is" condition is favoured
over "is-not" as it's more specific etc.
>From what I recall, in this particular case we should
be activating the other location as opposed to the
NoNet location. System locations are only supposed
to be activated if none of the conditions of any of
the other locations are met. However, in practice
I suspect that we probably apply NoNet when the
active NCU count goes to 0.
> In any case, I would say that a user-specified match should be
> ranked higher than a system-specified one; that would resovle
> the problem you're seeing. Other thoughts on this?
>
>
I'd agree with that. I'll see if I can figure out what's
going on here - I suspect a timed check is noticing
no IP NCUs are up, so we activate NoNet.
--
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.
