[squid-dev] [RFC] CRUSH peer selection method

[Loic Dachary]

Hi Amos,

On 04/19/2017 06:32 PM, Amos Jeffries wrote:
> [with maintainers hat on]
> 
> So your measurements are showing some slight differences, but not enough to make us immediately want to take it on. It also seems the use-case is not immediately clear and that implies we wont have much hope of community members jumping on it as a must-have and testing it for us.
> 
> You are always welcome to code it up yourself or to sponsor someone to do it.
> 
> Given the low-interest situation, I also request that whomever implements it be able to stick around and help with support (eg. bug fixes) for at least the first release series it gets published in. That usually means about 1 year, but with the present Squid-4 situation it may be a bit longer.

I understand that "changing the weight of a peer" is a marginal use case. And since noone complained about it so far, it's probably not worth trying to fix it.

Thanks for taking the time to discuss this, it was most interesting :-)

Cheers

-- 
Loïc Dachary, Artisan Logiciel Libre