Network Troubleshooting

[rafiki47]

Hello,
I have a Sunfire V210 with a 4 port network card. It was setup to the Public Network, and an Private. The Private no longer works. I get no lights in the back on the NIC. Could any one pass on some trouble shooting tips?

Thanks,

 

[cndcadams]

rafiki47;

to rule out the software bring system down to ok prompt.
Make sure cables are plugged in and run watch-net-all

This will check both loopback and packet transfers. You will get .... if the port is working. If it passes at low level then you can probably say your problem is in software configuration.


Are you talking about the 4 onboard ports?

If yes then check files;

more /etc/path_to_inst | grep bge (You should see 4 paths)

more /etc/hostname.bge* (you should have a config file for each configured port, the names or ip in the files should relate back to an entry in /etc/hosts)

run ifconfig -a (does this show you ports and if they are configured?)

Thanks

CA

 

[rafiki47]

CA,
Thanks for the help.

I ran watch-net-all and i got loopback failed errors. But after rebooting the server the lights on the 4port onbard NIC came back on.

When i do an ifconfig -a i only see one connection, so i will try and plumb that one and get back to you with the results.

Thanks,

 

[rafiki47]

Well not much Luck.
I have a SAN connected to a private network switch.It has an IP address of 192.168.128.101.

So on My Server i set bge1 - to 192.168.128.100

Both have the same subnet mask of 255.255.255.0

When i try to ping the address i get no reply.

Any ideas?

 

[cndcadams]

rafiki47;

did all the ports give you failed loopback or just one port when running the watch-net-all?

post your

more /etc/path_to_inst
more /etc/hostname.bge*
more /etc/hosts

ifconfig -a (output)

CA

 

[daFranze]

>> I have a SAN connected to a private network switch.It has an IP address of 192.168.128.101.

guessing it is a NAS... ;-)

run a snoop (-> man snoop) on that particular interface. It runs in promiscous mode, you should see every package on that subnet.

How is your 4NIC connected to the LAN? On switch? Do you have VLANs? What to the Network people say? did they enable "auto port shutdown" (don't know how they call it exactly), if a link fails to protect "attacs from inside" using MAC and IP of an "internal NIC".

Have you tried to use eg. a laptop/HUB/Switch and connect it via X-Link to one of the ports? Do you get a link there?

did you change any network parameters? fullduplex hardcoded? MAC Adress ("local-mac-address? true" change)?

Best Regards, Franz
--
UNIX System Manager from Munich, Germany

 

[rafiki47]

CA,
Heres the info you requested.



# more path_to_inst
#
# Caution! This file contains critical kernel state
#
"/pseudo" 0 "pseudo"
"/scsi_vhci" 0 "scsi_vhci"
"/options" 0 "options"
"/pci@1f,700000" 0 "pcisch"
"/pci@1f,700000/network@2" 0 "bge"
"/pci@1f,700000/network@2,1" 1 "bge"
"/pci@1e,600000" 1 "pcisch"
"/pci@1e,600000/ide@d" 0 "uata"
"/pci@1e,600000/ide@d/sd@0,0" 3 "sd"
"/pci@1e,600000/isa@7" 0 "ebus"
"/pci@1e,600000/isa@7/power@0,800" 0 "power"
"/pci@1e,600000/isa@7/rmc-comm@0,3e8" 0 "rmc_comm"
"/pci@1e,600000/isa@7/i2c@0,320" 0 "pcf8584"
"/pci@1e,600000/isa@7/i2c@0,320/motherboard-fru-prom@0,a2" 0 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/chassis-fru-prom@0,a8" 1 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/power-supply-fru-prom@0,b0" 2 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/dimm-spd@0,b6" 3 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/dimm-spd@0,b8" 4 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/dimm-spd@0,c6" 5 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/dimm-spd@0,c8" 6 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/nvram@0,50" 7 "seeprom"
"/pci@1e,600000/isa@7/i2c@0,320/gpio@0,44" 0 "pca9556"
"/pci@1e,600000/isa@7/i2c@0,320/gpio@0,46" 1 "pca9556"
"/pci@1e,600000/isa@7/i2c@0,320/gpio@0,68" 2 "pca9556"
"/pci@1e,600000/isa@7/i2c@0,320/gpio@0,70" 3 "pca9556"
"/pci@1e,600000/isa@7/serial@0,3f8" 0 "su"
"/pci@1e,600000/isa@7/serial@0,2e8" 1 "su"
"/pci@1e,600000/pmu@6" 0 "pmubus"
"/pci@1e,600000/pmu@6/gpio@80000000" 0 "pmugpio"
"/pci@1e,600000/pmu@6/i2c@0,0" 0 "smbus"
"/pci@1e,600000/usb@a" 0 "ohci"
"/pci@1e,600000/usb@a/mouse@1" 2 "hid"
"/pci@1e,600000/usb@a/keyboard@2" 0 "hid"
"/pci@1e,600000/usb@a/device@1" 0 "usb_mid"
"/pci@1e,600000/usb@a/device@1/keyboard@0" 1 "hid"
"/pci@1e,600000/usb@a/device@1/mouse@1" 3 "hid"
"/pci@1e,600000/usb@a/keyboard@1" 4 "hid"
"/pci@1e,600000/usb@a/mouse@2" 5 "hid"
"/memory-controller@0,0" 0 "mc-us3i"
"/memory-controller@1,0" 1 "mc-us3i"
"/pci@1c,600000" 2 "pcisch"
"/pci@1c,600000/scsi@2" 0 "glm"
"/pci@1c,600000/scsi@2/sd@0,0" 0 "sd"
"/pci@1c,600000/scsi@2/sd@1,0" 1 "sd"
"/pci@1c,600000/scsi@2,1" 1 "glm"
"/pci@1d,700000" 3 "pcisch"
"/pci@1d,700000/SUNW,XVR-100@1" 0 "pfb"
"/pci@1d,700000/network@2" 2 "bge"
"/pci@1d,700000/network@2,1" 3 "bge"
#
# more hostname.bge0
animal
# more hosts
#
# Internet host table
#
127.0.0.1 localhost
10.4.2.187 animal loghost
#
# ifconfig -a
lo0: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1
inet 127.0.0.1 netmask ff000000
bge0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
inet 10.4.2.187 netmask ffff0000 broadcast 10.4.255.255
ether 0:3:ba:e7:1:e1
bge1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
inet 192.168.128.100 netmask ffffff00 broadcast 192.168.128.255
ether 0:3:ba:e7:1:e2
#

 

[cndcadams]

rafiki47;

Well I am a little confused and maybe DaFranze can give us a hand.

Your ifconfig shows both devices as up but I do not show an entry in /etc/hosts for bge1 and you did not post your hostname.bge1 entry?

Your local_mac_address is set to true.

You path_to_inst shows your bge paths correctly.

"/pci@1f,700000/network@2" 0 "bge"
"/pci@1f,700000/network@2,1" 1 "bge"
"/pci@1d,700000/network@2" 2 "bge"
"/pci@1d,700000/network@2,1" 3 "bge"

Last thing you did not answer me on what ports fail the loopback from the watch-net-all command at ok prompt.
The below paths should pass both internal and external tests.

/pci@1f,700000/network@2 (this is port 0)
/pci@1f,700000/network@2,1 (This is port 1)

Let me know

Thanks

CA

 

[rafiki47]

CA,
The loopback failed on all four ports, and there isn't a hostname.bge1 file.



Thanks

 

[daFranze]

>> Well I am a little confused and maybe DaFranze can give us a hand.

sure, but with that output I can't say much more than you do.
With these infos you can just say: yes, the NIC is recognixed by the OS, yes it is configured corectly, the NICs are "UP" and "RUNNING". That's what you usually need to run a NIC.

Why does it not work? I don't know! I suggest to look at the system messages. I suggest to try a X-Link (With SX Cables this should not be a problem). Try from Port 0 to 1 or any other.
Did you read my post above...

Best Regards, Franz
--
System Manager (Solaris, HP-UX, Linux, some networking, some SAN)

 

[rafiki47]

Franz,
Ok i am trying to connect a Sunfire V210, and a StorageTek 6100 SAN. I think it is a SAN , but i have been wrong before. Anyway I am away from the office today, but i can run the suggestions tomorrow. I will get you guys feedback as soon as I can.

And Thanks for spending the time to help!

 

[dandan123]

You probably did a ifconfig command to set up bge1 which is why you do not have entries in /etc/host and /etc/hostname.bge1, you need these entries so the card is brought up during a reboot.

Check the status of the nic card using -

ndd -get /dev/bge1 link_status ( 0 = down 1 = up)
ndd -get /dev/bge1 link_speed
ndd -get /dev/bge1 link_dupex

 

[cndcadams]

rafiki47;

you should have a /etc/hostname.bge1 file with the name or ip which will relate to the /etc/hosts entry which you need to make if you want this port to be configured on reboot. AS dandan123 states you must be running the ifconfig command to setup the device which is a softset and will be lost after every reboot. Also if your ports are failing loopback I would say you have some hardware problem with the system board ports. If I get time tomorrow I will check how my v210 reacts to a watch-net-all. From what I remember every time I have used this command on other systems it would pass loopback unless the port was bad.

You can also try a ping -i 192.168.128.100 192.168.128.101

This will ping from you bge1 to your switch

Thanks

CA

 

[rafiki47]

OK guys,
Call me crazy but I can access the SAN now. So why does it work? I didnt change anything. But you are saying if i reboot it will stop working again?

 

[cndcadams]

Rafiki47;

I have never seen any other way to hard configure a port without creating a /etc/hostname.bge1 (create a name for the port) file, and make an entry for the port in /etc/hosts (name should be same as in hostname.bge1).

If you reboot without creating these files then yes you will have to rerun your ifconfig commands to get the port back up.

Thanks

CA

 

[daFranze]

well, it's not /dev/bge1 but /dev/bge, to get Infos from bge1 you need to set instance to "1" and afaik it's link_mode

Check the status of the nic card using -

ndd -set /dev/bge instance 1
ndd -get /dev/bge link_status ( 0 = down 1 = up)
ndd -get /dev/bge link_speed
ndd -get /dev/bge link_mode

Best Regards, Franz
--
System Manager (Solaris, HP-UX, Linux, some networking, some SAN)

 

[dandan123]

Franz,

The bge interface is different.

On the bge card it's /dev/bge1 , you do not have to set the instance like it's required on cards like the hme, and it's also link_duplex not link_mode.

 

[daFranze]

thanks dandan, lesson learned...

Best Regards, Franz
--
System Manager (Solaris, HP-UX, Linux, some networking, some SAN)

 

[dandan123]

Franz, I wasn't trying to teach you anything, you know far more than I do and I mean that seriously