Solaris 11 EA : DNS client ignoring secondary nameservers?
892055Oct 10 2011 — edited Jan 26 2012Greetings,
I'm testing Solaris 11 EA, and have encountered a DNS resolution problem.
My configuration is as follows (IP addresses have been modified) :
$ svccfg -s network/dns/client listprop config
config application
config/value_authorization astring solaris.smf.value.name-service.dns.client
config/domain astring my.domain
config/nameserver net_address 10.0.0.1 172.16.0.1
config/search astring my.domain
config/options astring timeout:2
The first nameserver is not accessible for the time being, and this configuration should mean that 2 seconds later, the second server should be queried. However, even though I can see both the query and the answer by snooping the interface, the OS ignores the answer and considers no servers could be reached :
$ host toto
;; connection timed out; no servers could be reached
While the sniffer dumps :
# snoop -rd net0 port 53
Using device net0 (promiscuous mode)
192.168.0.1 -> 10.0.0.1 DNS C toto.my.domain. Internet Addr ?
192.168.0.1 -> 172.16.0.1DNS C toto.my.domain. Internet Addr ?
172.16.0.1-> 192.168.0.1 DNS R toto.my.domain. Internet Addr 192.168.0.2
Is that a bug with the resolver, or am I overlooking something obvious?
Also, how can I declare multiple nameservers with svccfg? In my example above, the configuration was done at install time (through an AI profile), but when I try to change it manually, I can only declare a single address :
# svccfg -s network/dns/client
svc:/network/dns/client> setprop config/nameserver = 10.0.0.1 172.16.0.1
Unknown type "10.0.0.1".
I've tried various combinations, with or without quotes, with or without specifying the type as net_address, but no cookie.
Regards,
Daniel
Edited by: dpo on 10 oct. 2011 11:52