[4966] in testers
Re: erratic host lookup failures
daemon@ATHENA.MIT.EDU (Christopher D. Beland)
Tue Jul 10 05:34:07 2001
Message-Id: <200107100934.FAA06100@Press-Your-Luck.mit.edu>
To: Garry Zacheiss <zacheiss@MIT.EDU>
cc: testers@MIT.EDU
In-reply-to: The events that comprise the history of the universe.
Date: Tue, 10 Jul 2001 05:34:03 -0400
From: "Christopher D. Beland" <beland@MIT.EDU>
> This looks suspiciously like the failure documented in Sun Alert
> 26283, titled "Host Name Lookups Might Fail When Using DNS". The
> problem description does match what you describe (affects only
> Solaris 8, happens about 1% of the time).
Well, I am also experiencing the problem under Linux.
Actually, it just happened to me; I think it does indeed have
something to do with timeouts, and maybe also weird server
configurations or conditions. Here's what happened...
Netscape spewed a "could not find host or domain" error at me when I
attempted to access a file on www.kuro5shin.org. I played around a
bit, and I noticed that most of the name servers for that address time
out under (say) nslookup time out without giving the desired result.
In fact, there's only one name server that actually seems to work, and
it take some amount of querying to find it. Included below are some
select results from a lot of testing, following a lookup starting from
a gTLD server.
After I did all that, I flushed my DNS cache by
restarting athena-bind (which claims it is "Stopping Athena NTP" but
actually does the right thing). I did
nslookup -d www.kuro5shin.org
five times. The first time, it timed out after about 28 seconds and
didn't get the IP address. The second time, it succeeded after 9
seconds. Three more attempts produced a 9 second success, and two 28
second timeouts. (Restarting bind each time, obviously.) I did 5
trials with "host www.kuro5hin.org" on PYL. Sure enough, it took 5,
9, 32, 44, and 15 seconds, respectively, to get an answer.
I didn't notice any different results on whack-a-mole, which is a 9.0
Irix box, though I wasn't nearly as thorough there.
On an oooold Athena 8.2 SIPB-Linux box, I got four 9 second returns
and one 6 second return using the command "host www.kuro5hin.org". So
maybe something kooky is going on.
---
beland@Press-Your-Luck: nslookup -d www.kuro5hin.org paranoia.kuro5hin.org
;; res_querydomain(paranoia.kuro5hin.org, <Nil>, 1, 1)
;; res_query(paranoia.kuro5hin.org, 1, 1)
;; res_mkquery(0, paranoia.kuro5hin.org, 1, 1)
;; res_send()
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 59142
;; flags: rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; paranoia.kuro5hin.org, type = A, class = IN
;; Querying server (# 1) address = 127.0.0.1
;; got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 59142
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 4, ADDITIONAL: 4
;; paranoia.kuro5hin.org, type = A, class = IN
paranoia.kuro5hin.org. 1d21h15m16s IN A 63.111.177.29
kuro5hin.org. 10m2s IN NS dukat.thock.com.
kuro5hin.org. 10m2s IN NS kira.thock.com.
kuro5hin.org. 10m2s IN NS ns.intes.net.
kuro5hin.org. 10m2s IN NS paranoia.kuro5hin.org.
dukat.thock.com. 1d23h38m11s IN A 142.165.167.48
kira.thock.com. 1d16h50m10s IN A 142.165.167.47
ns.intes.net. 1d23h39m29s IN A 63.111.177.5
paranoia.kuro5hin.org. 1d21h15m16s IN A 63.111.177.29
;; res_mkquery(0, 29.177.111.63.in-addr.arpa, 1, 12)
timeout
timeout
The situation is the same under Irix...
beland@whack-a-mole: nslookup -d www.kuro5hin.org paranoia.kuro5hin.org
;; res_querydomain(paranoia.kuro5hin.org, <Nil>, 1, 1)
;; res_query(paranoia.kuro5hin.org, 1, 1)
;; res_mkquery(0, paranoia.kuro5hin.org, 1, 1)
;; res_send()
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 1276
;; flags: rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; paranoia.kuro5hin.org, type = A, class = IN
;; Querying server (# 1) address = 127.0.0.1
;; got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 1276
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 4, ADDITIONAL: 4
;; paranoia.kuro5hin.org, type = A, class = IN
paranoia.kuro5hin.org. 1d14h58m34s IN A 63.111.177.29
kuro5hin.org. 23m55s IN NS dukat.thock.com.
kuro5hin.org. 23m55s IN NS kira.thock.com.
kuro5hin.org. 23m55s IN NS ns.intes.net.
kuro5hin.org. 23m55s IN NS paranoia.kuro5hin.org.
dukat.thock.com. 1d23h51m23s IN A 142.165.167.48
kira.thock.com. 1d19h11m11s IN A 142.165.167.47
ns.intes.net. 1d23h52m50s IN A 63.111.177.5
paranoia.kuro5hin.org. 1d14h58m34s IN A 63.111.177.29
;; res_mkquery(0, 29.177.111.63.in-addr.arpa, 1, 12)
timeout
timeout
beland@Press-Your-Luck: dig @dukat.thock.com www.kuro5hin.org T_ANY C_ANY
; <<>> DiG 8.1 <<>> @dukat.thock.com www.kuro5hin.org T_ANY C_ANY
; (1 server found)
;; res options: init recurs defnam dnsrch
[hangs]
(5:02 ~) beland@Press-Your-Luck: dig @kira.thock.com www.kuro5hin.org T_ANY C_ANY
; <<>> DiG 8.1 <<>> @kira.thock.com www.kuro5hin.org T_ANY C_ANY
; (1 server found)
;; res options: init recurs defnam dnsrch
[hangs]
(5:03 ~) beland@Press-Your-Luck: dig @kira.thock.com www.kuro5hin.org
; <<>> DiG 8.1 <<>> @kira.thock.com www.kuro5hin.org
; (1 server found)
;; res options: init recurs defnam dnsrch
;; got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 10
;; flags: qr aa rd; QUERY: 1, ANSWER: 1, AUTHORITY: 4, ADDITIONAL: 3
;; QUERY SECTION:
;; www.kuro5hin.org, type = A, class = IN
;; ANSWER SECTION:
www.kuro5hin.org. 1H IN CNAME kuro5hin.org.
;; AUTHORITY SECTION:
kuro5hin.org. 1H IN NS dukat.thock.com.
kuro5hin.org. 1H IN NS kira.thock.com.
kuro5hin.org. 1H IN NS ns.intes.net.
kuro5hin.org. 1H IN NS paranoia.kuro5hin.org.
;; ADDITIONAL SECTION:
dukat.thock.com. 10M IN A 142.165.167.48
kira.thock.com. 10M IN A 142.165.167.47
paranoia.kuro5hin.org. 1H IN A 63.111.177.29
;; Total query time: 138 msec
;; FROM: Press-Your-Luck.mit.edu to SERVER: kira.thock.com 142.165.167.47
;; WHEN: Tue Jul 10 05:03:43 2001
;; MSG SIZE sent: 34 rcvd: 193
beland@Press-Your-Luck: dig @paranoia.kuro5hin.org www.kuro5hin.org
; <<>> DiG 8.1 <<>> @paranoia.kuro5hin.org www.kuro5hin.org
; (1 server found)
;; res options: init recurs defnam dnsrch
[hangs]
----paranoia.kuro5hin.org PING Statistics----
223 packets transmitted, 0 packets received, 100.0% packet loss
----dukat.thock.com PING Statistics----
4 packets transmitted, 4 packets received, 0.0% packet loss
round-trip min/avg/max = 163.429/163.985/164.587 ms
But:
beland@Press-Your-Luck: nslookup -d www.kuro5hin.org paranoia.kuro5hin.org
;; res_querydomain(paranoia.kuro5hin.org, <Nil>, 1, 1)
;; res_query(paranoia.kuro5hin.org, 1, 1)
;; res_mkquery(0, paranoia.kuro5hin.org, 1, 1)
;; res_send()
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 7709
;; flags: rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; paranoia.kuro5hin.org, type = A, class = IN
;; Querying server (# 1) address = 127.0.0.1
;; got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 7709
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 4, ADDITIONAL: 4
;; paranoia.kuro5hin.org, type = A, class = IN
paranoia.kuro5hin.org. 1d12h39m55s IN A 63.111.177.29
kuro5hin.org. 2m4s IN NS dukat.thock.com.
kuro5hin.org. 2m4s IN NS kira.thock.com.
kuro5hin.org. 2m4s IN NS ns.intes.net.
kuro5hin.org. 2m4s IN NS paranoia.kuro5hin.org.
dukat.thock.com. 1d23h24m52s IN A 142.165.167.48
kira.thock.com. 1d16h38m41s IN A 142.165.167.47
ns.intes.net. 1d23h30m47s IN A 63.111.177.5
paranoia.kuro5hin.org. 1d12h39m55s IN A 63.111.177.29
;; res_mkquery(0, 29.177.111.63.in-addr.arpa, 1, 12)
timeout
Finally, this command returns an instant answer:
beland@Press-Your-Luck: nslookup -d www.kuro5hin.org ns.intes.net
;; res_querydomain(ns.intes.net, <Nil>, 1, 1)
;; res_query(ns.intes.net, 1, 1)
;; res_mkquery(0, ns.intes.net, 1, 1)
;; res_send()
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 43144
;; flags: rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; ns.intes.net, type = A, class = IN
;; Querying server (# 1) address = 127.0.0.1
;; got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 43144
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 3, ADDITIONAL: 3
;; ns.intes.net, type = A, class = IN
ns.intes.net. 1d23h35m22s IN A 63.111.177.5
intes.NET. 1d23h35m22s IN NS NS2.intes.NET.
intes.NET. 1d23h35m22s IN NS NS1.intes.NET.
intes.NET. 1d23h35m22s IN NS NS3.intes.NET.
NS2.intes.NET. 1d23h35m22s IN A 151.201.22.73
NS1.intes.NET. 1d23h35m22s IN A 63.111.177.4
NS3.intes.NET. 1d23h35m22s IN A 63.111.177.6
;; res_mkquery(0, 5.177.111.63.in-addr.arpa, 1, 12)
------------
Got answer:
HEADER:
opcode = QUERY, id = 43145, rcode = NOERROR
header flags: response, auth. answer, want recursion, recursion avail.
questions = 1, answers = 1, authority records = 3, additional = 3
QUESTIONS:
5.177.111.63.in-addr.arpa, type = PTR, class = IN
ANSWERS:
-> 5.177.111.63.in-addr.arpa
name = mail-hub.intes.net
ttl = 86400 (1D)
AUTHORITY RECORDS:
-> 177.111.63.in-addr.arpa
nameserver = ns1.intes.net
ttl = 86400 (1D)
-> 177.111.63.in-addr.arpa
nameserver = ns2.intes.net
ttl = 86400 (1D)
-> 177.111.63.in-addr.arpa
nameserver = ns3.intes.net
ttl = 86400 (1D)
ADDITIONAL RECORDS:
-> ns1.intes.net
internet address = 63.111.177.4
ttl = 86400 (1D)
-> ns2.intes.net
internet address = 151.201.22.73
ttl = 86400 (1D)
-> ns3.intes.net
internet address = 63.111.177.6
ttl = 86400 (1D)
------------
Server: mail-hub.intes.net
Address: 63.111.177.5
;; res_mkquery(0, www.kuro5hin.org, 1, 1)
------------
Got answer:
HEADER:
opcode = QUERY, id = 43146, rcode = NOERROR
header flags: response, want recursion, recursion avail.
questions = 1, answers = 2, authority records = 4, additional = 4
QUESTIONS:
www.kuro5hin.org, type = A, class = IN
ANSWERS:
-> www.kuro5hin.org
canonical name = kuro5hin.org
ttl = 3293 (54m53s)
-> kuro5hin.org
internet address = 209.208.150.45
ttl = 1192 (19m52s)
AUTHORITY RECORDS:
-> kuro5hin.org
nameserver = dukat.thock.com
ttl = 1192 (19m52s)
-> kuro5hin.org
nameserver = kira.thock.com
ttl = 1192 (19m52s)
-> kuro5hin.org
nameserver = ns.intes.net
ttl = 1192 (19m52s)
-> kuro5hin.org
nameserver = paranoia.kuro5hin.org
ttl = 1192 (19m52s)
ADDITIONAL RECORDS:
-> dukat.thock.com
internet address = 142.165.167.48
ttl = 162024 (1d21h24s)
-> kira.thock.com
internet address = 142.165.167.47
ttl = 162024 (1d21h24s)
-> ns.intes.net
internet address = 63.111.177.5
ttl = 86400 (1D)
-> paranoia.kuro5hin.org
internet address = 63.111.177.29
ttl = 153920 (1d18h45m20s)
------------
Non-authoritative answer:
Name: kuro5hin.org
Address: 209.208.150.45
Aliases: www.kuro5hin.org
::whew::
-Beland
===============================================================
Christopher Beland - http://web.mit.edu/beland/www/contact.html
MIT STS/Course 6 (EECS) - MIT Athena User Interface Project
The Talk of MIT ** http://web.mit.edu/beland/talk/talk.html
===============================================================
