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

Re: glibc's getaddrinfo() sort order



Anthony Towns wrote:
> So if getaddrinfo() has always behaved in this way, I don't see a great
> deal of justification in changing it. The bug log indicated that there
> were pre-rfc implementations of getaddrinfo() that behaved more like
> gethostbyname() at least wrt round-robin DNS; but I've got no way of
> verifying that.

This is a locally compiled wget because I wanted to use the same version
that's in unstable. In particular, it's configured with --enable-ipv6,
so it will use getaddrinfo. gluck is still using debian 3.1, so libc6
2.3.2:

joeyh@gluck:~/tmp/wget-1.10.2>ltrace src/wget http://http.us.debian.org/ 2>&1 |grep getaddr                        
getaddrinfo("http.us.debian.org", NULL, 0xbfb5fdd0, 0xbfb5fdcc) = 0
joeyh@gluck:~/tmp/wget-1.10.2>for x in 1 2 3 4 5; do src/wget http://http.us.debian.org/ 2>&1 |grep Connecting; done 
Connecting to http.us.debian.org|64.50.238.52|:80... connected.
Connecting to http.us.debian.org|35.9.37.225|:80... connected.
Connecting to http.us.debian.org|128.101.240.212|:80... connected.
Connecting to http.us.debian.org|64.50.238.52|:80... connected.
Connecting to http.us.debian.org|128.101.240.212|:80... connected.

Now here is the same wget running with unstable's libc, 2.6.1:

joey@kodama:~>scp gluck.debian.org:~/tmp/wget-1.10.2/src/wget .
wget                                          100%  227KB 226.8KB/s   00:01    
joey@kodama:~>for x in 1 2 3 4 5; do ./wget http://http.us.debian.org/ 2>&1 |grep Connecting; done
Connecting to http.us.debian.org|204.152.191.7|:80... connected.
Connecting to http.us.debian.org|204.152.191.7|:80... connected.
Connecting to http.us.debian.org|204.152.191.7|:80... connected.
Connecting to http.us.debian.org|204.152.191.7|:80... connected.
Connecting to http.us.debian.org|204.152.191.7|:80... connected.

> AFAICS round-robin DNS isn't an Internet standard either, for that matter,
> which seems odd to me. There's a Microsoft page that references rfc1794
> for round-robin load balancing, but it doesn't seem entirely on point, and
> is just a memo anyway.

Well you know what they say about implementing TCP/IP to the standards,
right? You get something that doesn't work on the actual internet.
 
-- 
see shy jo

Attachment: signature.asc
Description: Digital signature


Reply to: