[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Again ... Re: security.debian.org

> Matus UHLAR - fantomas wrote:
> > However this does not _seem_ to be enough, and I guess it's due to "feature"
> > of glibc, that sorts IP addresses from numerically lowest one to numerically
> > highest one (so it spoils attempts for DNS load-balancing), so if any program
> > is the order of addresses returned via gethostbyname() or getnameinfo(), it
> > tries them always in the following order:
> > 
> > % getent hosts security.debian.org
> > security.debian.org
> >  security.debian.org
> > security.debian.org
> > 
> > so no wonder _if_ is loaded more than other servers.
> > many times I have to re-try update/upgrade to ask other servers and
> > seems to be the slowest for me.

On 29.03.07 13:42, virendra rode // wrote:
> I was under the impression that why netselect (picks closets and fastest
> server) was implemented.

I'm sure it was not implemented to avoid lame feature of glibc. You can't
run netselect before any of e.g. DNS lookup or TCP connect to a server who
has more than one address.
Matus UHLAR - fantomas, uhlar@fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
He who laughs last thinks slowest. 

Reply to: