Hello all,
I have a sun server SunOS PNR1RAD01 5.10 Generic_144488-09 sun4v sparc SUNW,Netra-T5220 where i have an interface e100g0.
This was earlier in running state but now this doesnt show running state. I tried to reset the interface, checked the cable but it did not help.
Can nyone pls help me on this.
e1000g0: flags=1000803<UP,BROADCAST,MULTICAST,IPv4> mtu 1500 index 2
inet 10.200.2.73 netmask fffffff8 broadcast 10.200.2.79
ether 0:21:28:bd:55:3e
What changes did you make to your system? Did you edit any files such as /etc/hostname.e1000g0?
nothing was done on the server,
but yes one observation, though it doesn't show running state but the ip defined on this interface is well reachable from another node. This eliminated my suspicion on the cable part. I also restarted the whole server as well but no luck.
Can someone advice further ??
So what do you mean by not showing a "running state"?
Your quote above shows the interface as "up"
Your output shows its already up and running and you told ip is pingable then what else you are looking for?
If still any problem, just unplumb and plumb again and do all the configuration again.
see the other interface o/p:
ossuser@PNR1RAD01 # ifconfig -a
lo0: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1
inet 127.0.0.1 netmask ff000000
e1000g0: flags=1000803<UP,BROADCAST,MULTICAST,IPv4> mtu 1500 index 2
inet 10.200.2.73 netmask fffffff8 broadcast 10.200.2.79
e1000g1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
inet 10.200.3.58 netmask ffffffc0 broadcast 10.200.3.63
And one more observation:
root@PNR1RAD01 # dladm show-dev
e1000g0 link: down speed: 0 Mbps duplex: half
e1000g1 link: up speed: 1000 Mbps duplex: full
e1000g2 link: unknown speed: 0 Mbps duplex: half
e1000g3 link: unknown speed: 0 Mbps duplex: half
This link shows down in this command o/p although in ifconfig o/p it shows up but noot the running state as e1000g1 interface.
Please post output of:
cat /etc/hostname.e1000g0
and
cat /etc/hostname.e1000g1
Do the cables from e1000g0 and e1000g1 go to the same network switch?
If so, what make/model of switch? Does it understand multipaths?
ossuser@PNR1RAD01 # cat hostname.e1000g0
PNR1RAD01
ossuser@PNR1RAD01 # cat hostname.e1000g1
10.200.3.58
ossuser@PNR1RAD01 #
yes these cables go to the same switch, pls note this was working before but somehow this suddenly went down.
Normally, a /etc/hostname.<interface name> file contains just a hostname (as you have in your hostname.e1000g0 file). The host names provided in these files is looked-up in /etc/hosts to tell Unix what ip address to use for that interface.
Anyway, you have different ip subnet addresses on these interfaces which, depending on how clever your network switch is, might confuse the hell out of the switch and cause it to "down" one of the interfaces.
You could provide different hostnames in each "hostname" file and different ip addresses for these in /etc/hosts.
Alternatively, if your switch will understand, you could set up an aggregate interface (Google that "aggr1") to combine two or more interfaces to look like one and load share via a clever switch.
root@PNR1RAD01 # cat /etc/hosts
#
# Internet host table
#
127.0.0.1 localhost
::1 localhost
10.200.2.73 PNR1RAD01 loghost
10.200.3.58 PNR1RAD01
yes its defined as above, but am still not able to resolve this issue, gave a soft reboot to the server as well but it did not help to resolve the issue.
With a /etc/hosts containing the same hostname twice, only the first entry will ever be used when the system searches to resolve a hostname.
That's why you have an ip address in the second hostname file.
This is very unusual.
You are definitely saying that both network interfaces were working a short while ago? Has the network switch been changed recently?
Or, has there been another change on the network that might have created a loop that the switch can see? This could be just a cable being moved or something elsewhere on your network.
I reckon that network conditions (loops???) are causing your switch to drop one of the Solaris interfaces.
Hopefully, I can get a second opinion on this forum from another Solaris networking expert. I can't think of much else it can be.
---------- Post updated at 08:55 PM ---------- Previous update was at 08:15 PM ----------
Here's an idea..........
If you've got a spare switch put each interface cable into different switches. Connect just a PC or something to the spare switch too.
See if you can ping the respective ip address from each "separate" network and also see if both interfaces stay up.
If the interfaces stay up then that pretty much eliminates the Solaris box as being the problem.