Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

DNS issues after 10.6.5 update

Since updating to 10.6.5 (via the combo update), I experience strange DNS problems on my laptop.

After working fine for some time, suddenly no DNS names are resolved any longer.
Looks like mDNSResponder simply died.

- This happens simultaneously in Firefox, Safari & even from the Terminal.
- It happens both on ethernet, and on wifi.
- I can still Ping my DNS servers, and I can ping my machine from remote as well.
- Everything where I can still remember the IP address seem to work fine.
- Problem happened in different networks
- Nothing obvious in the console messages

It really looks like a pure DNS problem, not a network issue.
Even the Windows emulation I sometimes run on that machine keeps working and can still resolve DNS names correctly.

I repaired permissions, reinstalled the update, cleared the DNS cache... the usual.

Until now, nothing solved the problem. I need to reboot, to get dns lookup back.
This will then work for some time (in the order of hours) and suddenly I'm stuck again.

cheers

Martin

MacBook Pro 15" i7, Mac OS X (10.6.5)

Posted on Nov 14, 2010 2:48 PM

Reply
130 replies

Nov 20, 2010 6:43 PM in response to Martin Jaekel1

In my case I'm always using the computer when the problem happens... suddenly hostnames stop being resolved. So I don't think it's a sleep issue for me. I've also been unable to reset mDNSResponder by either resetting the network connection or changing the network connect. Once mDNSResponder craps out turning the airport off/on or switch from wireless to wired has no affect.

Nov 22, 2010 12:36 AM in response to darren2002

Thanks for the tip on changing CNAMEs to A records! This works!!

It would seem that for some reason Mac os X 10.6.5 will not accept domain aliases that resolve to out of zone.

I guess this makes sense given the number of DNS intercepting viruses/trojans etc that are out there on PCs - maybe Apple are pre-empting something!

Nov 24, 2010 11:58 AM in response to Martin Jaekel1

So it appears that nobody has solved the problem as of yet.

The one thing to add is that this seems to have something to do
with /usr/sbin/awascd. I don't really know what this process is for
but it appears to have something to do with name lookups via standard
gethostbyaddr etc.

Currently I see 2 situations: first name lookup completely hang and the calling
process is stuck i.e. ping foo.example.com.
second name lookup fail with a host not found message.
But the whole time nslookup / host / dig works just fine, so it's not an issue of talking
to the DNS server.

awacsd seems to be something new in 10.6.5 since I don't find it on my 10.6.4 system (which is working just fine).

Anybody know how to debug awascd

-Russell

Nov 24, 2010 3:38 PM in response to andyBall_uk

For my part, I have Back to my mac not enabled, and did not see the awascd process.
I can't confirm that it is linked. But I will keep an eye for this now.

The problem continues to hit me in a completely random (?) way. Sometimes I go a whole day without reboot, sometimes I have it a few minutes after the last reboot. I simply can't find a pattern of the conditions under which this bug is triggered (like network activity, programs used, web sites visited, plug-ins loaded...)

Out of the blue, the DNS name resolving fails system wide.

Nov 25, 2010 4:09 AM in response to darren2002

I've been hit with this problem, and I think I can see what's going wrong.

I have an internal Active Directory domain which is used for DNS, which is on a .local (we'll call it mydomain.local)
We also have another domain (we'll call it otherdomain.com) which is used for externally accessable resources, we have an internal DNS server with otherdomain.com as an ctive zone with CNAME records pointing to .mydomain.local hosts.

This is a common setup, I work with a lot of businesses and this is a very standard way to do things, so you can give users 1 hostname for access to a service and internally it points to the internal IP and externally it points to your forwarded external IP.

OK, enough history.

I put mDNSResponder into debug mode as per it's man page entry

I then 'pinged' an otherdomain.com address which resolves as a CNAME to a mydomain.local address

Here is the logs:

Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceCreateConnection START
Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Warning>: 14: Error socket 31 created 00000000 00000001
Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceQueryRecord(5000, 0, mx2.otherdomain.com., Addr) START
Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Info>: GetServerForQuestion: 0000000101006C88 DNS server 192.168.2.222:53 (Penalty Time Left 0) (Scope None:0000000000000000) found for name mx2.otherdomain.com. (Addr)
Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Info>: mDNS StartQueryinternal: question 0000000101006C88 mx2.otherdomain.com. (Addr), DNS Server 192.168.2.222:53
Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Warning>: 14: Error socket 31 closed 00000000 00000001 (0)
Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceQueryRecord(mx2.otherdomain.com., Addr) ADD 23 mx2.otherdomain.com. CNAME app2.mydomain.local.
Thu Nov 25 11:42:36 smlt8001 mDNSResponder[31] <Info>: AnswerCurrentQuestionWithResourceRecord: 0000000101006C88 mx2.otherdomain.com. (Addr) following CNAME referral 0 for 23 mx2.otherdomain.com. CNAME app2.mydomain.local.
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Warning>: 14: Cancel 00000000 00000001
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceQueryRecord(app2.mydomain.local., Addr) STOP
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Warning>: 14: Error socket 31 created 00000000 00000002
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceQueryRecord(5000, 0, mx2.otherdomain.com.mydomain.local., Addr) START
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceQueryRecord(mx2.otherdomain.com.mydomain.local., Addr) unicast
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Info>: GetServerForQuestion: 0000000100837D78 DNS server 192.168.2.222:53 (Penalty Time Left 0) (Scope None:0000000000000000) found for name mx2.otherdomain.com.mydomain.local. (Addr)
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Info>: mDNS StartQueryinternal: question 0000000100837D78 mx2.otherdomain.com.mydomain.local. (Addr), DNS Server 192.168.2.222:53
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Warning>: 14: Error socket 31 closed 00000000 00000002 (0)
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Info>: mDNSCoreReceiveResponse ignoring 58 mydomain.local. SOA dc.mydomain.local. hostmaster. 47683 900 600 86400 900
Thu Nov 25 11:43:06 smlt8001 mDNSResponder[31] <Info>: Skipping check to see if we need to generate a negative cache entry for mx2.otherdomain.com.mydomain.local. (Addr)
Thu Nov 25 11:43:07 smlt8001 mDNSResponder[31] <Info>: mDNSCoreReceiveResponse ignoring 58 mydomain.local. SOA dc.mydomain.local. hostmaster. 47683 900 600 86400 900
Thu Nov 25 11:43:07 smlt8001 mDNSResponder[31] <Info>: Skipping check to see if we need to generate a negative cache entry for mx2.otherdomain.com.mydomain.local. (Addr)
Thu Nov 25 11:43:08 smlt8001 mDNSResponder[31] <Warning>: 14: Cancel 00000000 00000002
Thu Nov 25 11:43:08 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceQueryRecord(mx2.otherdomain.com.mydomain.local., Addr) STOP
Thu Nov 25 11:43:08 smlt8001 mDNSResponder[31] <Warning>: 14: DNSServiceCreateConnection STOP

This is messed up.

It's looking up mx2.otherdomain.com and getting a CNAME of app2.mydomain.local.
It's then trying to look up mx2.otherdomain.com.mydomain.local

Why?

At first I thought it was not handling the '.' on the end of the CNAME and treating it as a relative name rather than a FQDN, but that's not it, because that would resolve to app2.mydomain.local.otherdomain.com NOT mx2.otherdomain.com.mydomain.local

Something is seriously hosed in the part of mDNSResponder in 10.6.5 when it handles a CNAME to a .local domain

Steven

Message was edited by: SBlackery

Nov 25, 2010 6:47 AM in response to Martin Jaekel1

I have same problems since update to 10.6.5.
Process of mDNSresponder was crashed and reported about crashreport at /Library/Logs/CrashReporter.

List on CrashReport:
------
admin-no-macbook:log root# cd /Library/Logs/CrashReporter/
admin-no-macbook:CrashReporter root# ls
mDNSResponder 2010-11-14-235729localhost.crash
mDNSResponder 2010-11-17-142126localhost.crash
mDNSResponder 2010-11-18-090429localhost.crash
mDNSResponder 2010-11-18-125943localhost.crash
mDNSResponder 2010-11-18-210321localhost.crash
mDNSResponder 2010-11-20-133039localhost.crash
mDNSResponder 2010-11-22-071250localhost.crash
mDNSResponder 2010-11-23-084515localhost.crash
mDNSResponder 2010-11-24-120604localhost.crash
mDNSResponder 2010-11-25-193637localhost.crash
admin-no-macbook:CrashReporter root#

------

When crash process of mDNSResponder, segfault the process and record system.log:
Nov 25 19:36:28 admin-no-macbook mDNSResponder[4570]: * process 4570 exceeded 500 log message per second limit - remaining messages this second discarded *
Nov 25 19:36:30 admin-no-macbook mDNSResponder[4570]: * process 4570 exceeded 500 log message per second limit - remaining messages this second discarded *
Nov 25 19:36:31 admin-no-macbook mDNSResponder[4570]: * process 4570 exceeded 500 log message per second limit - remaining messages this second discarded *
Nov 25 19:36:36 admin-no-macbook com.apple.launchd[1] (com.apple.mDNSResponder[4570]): Job appears to have crashed: Segmentation fault
Nov 25 19:36:37 admin-no-macbook com.apple.ReportCrash.Root[6323]: 2010-11-25 19:36:37.721 ReportCrash[6323:2a03] Saved crash report for mDNSResponder[4570] version ??? (???) to /Library/Logs/DiagnosticReports/mDNSResponder 2010-11-25-193637localhost.crash

Under happens crashing process of mDNSResponder, I still can execute that:
- ping defaultgateway and DNS server, by IP Address.
- lookup hosts by nslookup (eg www.google.com, www.apple.com )
- host command (eg host admin-no-macbook.local has address 192.168.3.100)
But I cannot execute that.
- cannot browse internet (Firefox, Safari ...)
- "telnet www.google.com 80" is not responsed. (nodename nor servname provided, or not known)
- opening system Preference - share is not responsed.

I hove nothing solved this problem, too.

Nov 26, 2010 3:37 PM in response to Martin Jaekel1

This works for me. When DNS craps out on OSX, I switch over to XP (under Parallels Desktop, but it will probably work the same under Fusion) and if I select the Default Adapter for Bridged Networking in Parallels Prefs then DNS still works in all the XP web browsers. It's not exactly an ideal solution, but it's not that bad either because it turns out that Firefox on XP is much faster than on OSX. Firefox has become a real dog on OSX.

DNS issues after 10.6.5 update

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple ID.