[olsr-dev] Seems like a bug...

[Sven-Ola Tuecke]

Thomas,

is it possible to change that, i.e. have a shorter valtime than 
win*interval? Oh yes - what for: Like to have up to 16 minutes measurement 
window for more precision but don't want a dead link in the tables for that 
long time. Do I miss something hiere, or is it just a matter of discarding 
that plausicheck?

Another elektra-suggestion: TCs twice as fast as Hellos for more routing 
stability in case of a lossy link...

LG Sven-Ola

"Thomas Lopatic" <(spam-protected)> schrieb im Newsbeitrag 
news:(spam-protected)
> Hey Elektra.
>
> [...]
>
>> WARNING: wlan0 HELLO validity time set to 1000.0 seconds!
>> WARNING: eth0 HELLO validity time set to 1000.0 seconds!
>>
>> That's the relevant part in my olsrd.conf:
>>
>> LinkQualityWinSize      100
>>
>>     HelloInterval 10.0
>>     HelloValidityTime   100.0
>
> [...]
>
> It's a feature!! With a LQ window size of 100 and a HELLO interval of 10
> seconds olsrd now requires a minimal HELLO validity time of 100 x 10
> seconds = 1000 seconds. If you specify a shorter validity time, olsrd
> automatically ups this to 1000 seconds. That's what the warning is
> about. With a window size of 100 we do not want to lose a link after 100
> seconds, i.e. after only 10 missed HELLOs. Or am I missing something here?
>
> Hmmm. Maybe the warning message is a bit misleading as it suggests that
> the user has set the validity time to 1000 seconds.
>
> -Thomas
>
> _______________________________________________
> olsr-dev mailing list
> (spam-protected)
> https://www.olsr.org/mailman/listinfo/olsr-dev