http://defect.opensolaris.org/bz/show_bug.cgi?id=9340
--- Comment #3 from amaguire <alan.maguire at sun.com> 2009-06-25 11:13:42 ---
(In reply to comment #2)
> I've reproduced this, and I'm seeing the same problem as you. The crux
> of the issue seems to be name resolution, as I see autofs log messages saying
> "automountd: server localhost not responding", and "automountd: self_check:
> unknown host: localhost" daemon.debug messages, which show the reason
> localhost
> isn't loopback mounting is name resolution.
>
> I also see the same things you are - "getent hosts localhost" works, but
> pinging it fails with "unknown host". Interestingly, "telnet localhost 7"
> (with
> echo:stream enabled) succeeds in resolving localhost to 127.0.0.1 and
> connecting, so I'm going to try and determine how telnet's name resolution
> differs from that used by ping and friends. This may point us at the real
> source of the problem.
>
> What I find weird is I can't see how anything location application does could
> cause these problems. I'll keep digging though.
I thought I'd gotten to the bottom of this, and then I retested with the
latest nwamd, and this problem has just gone away. Weird. Anyway, give it a try
with the latest builds (which are in flight at the moment, should be ready late
tonight), and let me know if it works for you.
Here's what I discovered in the meantime. A lot of applications - ping being
one of them - use getaddrinfo(3SOCKET) to carry out name-to-address mappings.
These applications have in common that they specify the AI_ADDRCONFIG flag.
>From the manpage:
If the AI_ADDRCONFIG flag is specified, IPv4 addresses are
returned only if an IPv4 address is configured on the local
system, and IPv6 addresses are returned only if an IPv6
address is configured on the local system. For this case,
the loopback address is not considered to be as valid as a
configured address. For example, when using the DNS, a query
for AAAA records should occur only if the node has at least
one IPv6 address configured (other than IPv6 loopback) and a
query for A records should occur only if the node has at
least one IPv4 address configured (other than the IPv4 loop-
back).
So I discovered that if the NoNet location is active and no v4 interfaces are
plumbed (as will be the case if WiFi can't find an AP and wired isn't plugged
in), we get these failures. However, just plumbing an interface makes them go
away.
Why they've gone away for the latest bits I'm running with, I have no idea. But
my testing - which failed before exactly like yours - now shows that "ping
localhost" still fails while autofs succeeds in loopback mounting /home/darren,
suggesting that the ping name resolution and autofs failures are different
issues.
--
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
