This is the mail archive of the libc-help@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Question about getaddrinfo taking 20 seconds


On 16/07/12 20:07, Yann Droneaud wrote:
Hi,

Le lundi 16 juillet 2012 Ã 12:53 +0200, Eugen Dedu a Ãcrit :

I have a dns-related question. The following program shows here 20 seconds, which means that getaddrinfo with AI_CANONNAME takes 20 seconds. Is it normal?

[...]


const char *name = "stun.ekiga.net";

[...]


Request with dig "stun.ekiga.net" A, returns:
  stun.ekiga.net.		2385	IN	CNAME	stun.voipbuster.com.
  stun.voipbuster.com.	516	IN	A	77.72.169.158
  stun.voipbuster.com.	516	IN	A	77.72.169.156
  stun.voipbuster.com.	516	IN	A	77.72.169.154
  stun.voipbuster.com.	516	IN	A	77.72.169.152

When trying a retrieve reverse PTR records for
77.72.169.158, 77.72.169.156, 77.72.169.154, 77.72.169.152
with dig -x, I'm getting either "SERVFAIL" or timeout errors.

So, the latencies are probably related to the reverse PTR which seems to
be misconfigured.

In fact, according to some DNS server/resolver log, the exact error is:

Jul 16 20:01:45 quoi named[6493]: error (unexpected RCODE REFUSED)
resolving '152.169.72.77.in-addr.arpa/PTR/IN': 77.72.174.217#53
Jul 16 20:01:45 quoi named[6493]: error (host unreachable) resolving
'152.169.72.77.in-addr.arpa/PTR/IN': 77.72.175.217#53

I (or my friends) tried this program on five networks (one different machine per network), with the following times : 20, 20, 15, 5, and 0 seconds.


Why some of them do not timeout, while others do?

What can I do to solve this issue? Is there a workaround in the code? Should I contact the guys on stun.voipbuster.com to fix the issue?

--
Eugen


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]