[geeks] Solaris 10 networking is giving me fits
Micah R Ledbetter
vlack-lists at vlack.com
Mon Sep 18 01:12:36 CDT 2006
I've spent little time with Solaris before, and I'm sure I'm doing
something (probably something big) all wrong. But what?
I installed Solaris 10 as unconnected to the network. (By the way, is
there a convenient way to tell the Solaris installer "I want
everything, unless it depends on X11"? This is a Netra T1, and I just
hit the "Everything" button, since I didn't want to screw with it,
but having X11 on here is worth nothing.)
I was following Ben Rockwood's "An Accelerated Introduction to
Solaris" guides[0], and in the second of the series, he gives a nice
list of how solaris networking works (very compact, but still a
little long to include in this email, so not duplicated here, but
each step is shortened):
1. put hostname in /etc/nodename
2. Add the hostname and IP address to /etc/hosts
3. echo "192.168.1.0 255.255.255.0" > /etc/netmasks
4. put hostname in /etc/hostname.(interface)
5. edit /etc/resolv.conf
6. put router ip address in /etc/defaultrouter
7. svcadm restart svc:/network/physical (or reboot)
I have also done a 'cp /etc/nsswitch.dns /etc/nsswitch.conf'
I want to add the settings:
router 192.168.1.1
netmask 192.168.1.0 255.255.255.0
hostname foo
ip address 192.168.1.6
interface hme1
My cable is actually plugged into hme1 :)
When I follow these instructions, I get:
ip_arp_done: init failed
Failed to configure IPv4 interface(s): hme1
Sep 18 00:44:55 svc.startd[7]: svc:/network/physical:default: Method
"/lib/svc/m.
Sep 18 00:44:55 svc.startd[7]: network/physical:default
misconfigured: transitio(...)
/var/svc/log/network-physical:default.log shows
[ Sep 18 00:36:29 Enabled. ]
[ Sep 18 00:36:29 Executing start method ("/lib/svc/method/net-
physical") ]
[ Sep 18 00:36:29 Timeout override by svc.startd. Using infinite
timeout ]
ifconfig: SIOCSLIFNAME for ip: hme1: already exists
configuring IPv4 interfaces:ifconfig: setifflags: SIOCSLIFFLAGS:
hme1: Cannot ass
hme1.
add net default: gateway 192.168.1.1: Network is unreachable
[ Sep 18 00:36:29 Method "start" exited with status 96 ]
Very weird point: DHCP works... almost
1. When I get a DHCP address, it assigns me 192.168.0.100
(according to ifconfig), even though the router, a linksys WRT54G, is
handing out 192.168.1.* addresses.
2. DNS doesn't work, but pinging outside hosts by IP address (eg
the DNS servers 209.244.0.3 or 4.2.2.1) does work. Other DHCP clients
do DNS just fine.
3. If I add an /etc/resolv.conf while DHCP is enabled, it gives
the *same errors* as above, to STDERR and /var/svc/log/network-
physical:default.log.
4. Though I can ping other hosts, other hosts cannot ping me at
192.168.0.100.
Stuff I've tried:
1. Google :)
2. setenv local-mac-address? true. It wasn't before, but that
didn't solve it.
Questions:
1. Why do I get the SIOCSLIFFLAGS error?
2. Why does DHCP go all weird on me?
3. Why does 'ifconfig hme1 down' have no effect? 'ifconfig -a'
reports the same before and after. Is this something I should be
doing with svcadm?
Random thought: I like the svcadm stuff, but it would be nice if the
command had a list, where 'svcadm list' would list all of the
services it manages, and 'svcadm list net' would list all the
services which contain the string "net".
- Micah
[0] http://www.cuddletech.com/blog/pivot/entry.php?id=562 (part 1)
http://www.cuddletech.com/blog/pivot/entry.php?id=573 (part 2)
More information about the geeks
mailing list