Ultra E150 Solaris 2.5.1 Oracle server. was working fine, then started losing network connection. I can see in messages file that it was complaining about another device using its ip address. Though not any longer, though I wonder if that was a red herring. My IT guys have now re-reserved its ip address on the dhcp server. I have restarted it several times but it won't talk to the network. I tried changing its ip address to another reserved one, but Oracle complains, the ip must be embedded in that somewhere. Should the ethernet cable socket flash when connected on this hardware? it doesn't. I tried a different cable, direct into a different hub.
You should not have changed anything on the system. You should have disconnected and reconnect the cable or worst case you should have rebooted the system once since you found out the conflicting IP and isolating it
Thanks Incredible. Trying a different cable was the first thing tried, then a different port, with no success. After seeing the ip conflict message and fixing the reserved ip address, the server was restarted prior to changing anything else. It did not fix the problem. Then I only changed the IP address and its now changed back. The behaviour is still identical to what it was before. The messages file does not complain about an ip conflict since that was fixed. So I haven't actually isolated teh problem yet, though it may have been triggerred by the ip conflict.
Was this reverted back?
Yes the ip address is reserved as what it was originally on DHCP and I've reverted the etc/hosts file back to its original address.
where is your oracle communicating from? Am I correct to say that you are not able to ping your server remotely as before? No outputs, so Im having difficulty to figure out where the actual problem is
I am unable to ping to this server from the network, - nor from this server to the network.
Oracle is installed on this server.
Pls resolve this first
check if you can ping your gateway. netstat -rn ?
snoop -d hme0
always found this usefull. Can you see any traffic when you run this command?
... considering that the E150 is a desktop whose last ship date was in 1998, and it is running an OS whose last ship date was even earlier, I'd say there's a strong possibility the NIC went bad, or the driver got corrupted.
Hi Guys.
From netstat -rm I get a table of results, would they be useful to your analysis?
The snoop command returned nothing, it just seemed to stall.
---------- Post updated at 03:36 PM ---------- Previous update was at 03:35 PM ----------
Hi System Shock, can you help trouble shoot that line of enquiry?
netstat -rn output first pls.
netstat -rn output was this;
Destination Gateway Flags Ref Use Interface
127.0.0.1
---------- Post updated at 09:33 AM ---------- Previous update was at 09:30 AM ----------
netstat -rn output was this;
Destination Gateway Flags Ref Use Interface
127.0.0.1 127.0.0.1 UG 0 0 lo0
172.25.0.0 172.25.4.250 U 3 3 hme0
224.0.0.0 172.25.4.250 U 3 3 hme0
default 172.25.4.1 UG 0 0
Can you ping your gateway IP 172.25.4.1 ?
Hi Incredible, no is the short answer.
Any attempt to get through to the network results in the hme:0 link down problem.
---------- Post updated at 03:25 PM ---------- Previous update was at 02:30 PM ----------
In case it makes any difference, this is not a desktop, it is a floor standing server of some stature. Enterprise E150 server.
Ultra Enterprise 150 is listed under desktops in the Sun ESOL site, right under the Ultra 80... but really, that's irrelevant to troubleshoot the network connection - other than to determine hardware age which may be a factor on HW components mal-functioning.
Here are a couple things you can do to isolate the problem:
-
If it is a big self-standing server, it surely has more than 1 NIC. Configure another NIC and try your network connection on the other NIC.
-
or better yet, use another server/PC/laptop to test the network connection. Simply configure the server/PC/laptop with the E150's IP.
Obviously, in either case, you'd want to use the very same cable and switch port that are connected to hme0 in the E150.
Thanks - I will test the connection with another machine shortly. I guess this server was not a majorly expensive one as it only has one ethernet connection. Could it have a second NIc, connected to the same cable port? There is a picture of it here; Netra 150 / Ultra Enterprise 150
It probably doesn't look like a workstation to you, but that is a late 90's Sun workstation 
Basically an Ultra 80 with a stand. It would have one built in NIC, and probably an addon quad NIC if it was ordered that way. I believe the addon ports would come up as qfe0, qfe1, etc.. you can try plumbing qfe0, qfe1, etc and see if one plumbs...
Now that I think about it, I'd say it would be better to first try the network connection the E150 is using in another box. Just take a laptop, hardcode it with the IP info from the E150, have the network guys reserve the IP to the laptop's MAC (if that's what they are using to reserve the IP) then remove the cable from the E150's hme0 right and plug it right into the laptop. That will tell you if the network connection works or not.
If it works, then you know is something in the server.
excellent, will do. There is definitely no other interface when plumbing, unless you count le0?
Do one simple test. ensure that the IP on your system is plumbed up and is "UP". configure your laptop to an IP of the same subnet and see if you can ping. This will tell your if your Nic is good or not