Hi.

I have just installed solaris 8 on an intel platform using 3com fast ethernet interfaces. everything appears to work well except for communication past the network interface. I can ping the localhost, i can ping the local interfaces. But when I try to ping a host on the local network I get unreachable messages.

Looking at netstat shows that all packets appear to go the localhost loopback interface (lo0) but not to the interface (elxl0)

elxl0 does not appear to be sending/rec any packets. However when I ping another localhost no response. I even added the mac address of other hosts to the arp table but still no joy.

the interfaces are up and running so there is not a problem

Interestingly although i get unreachable messages when pinging another host I see network activity on both interfaces and the hub shows pinging traffic also between the 2 hosts.

I wonder if anybody can help

Report this post to a moderator

that you have unique IP addresses on each host? You also need to ensure that each host has the same netmask. Are your other boxes PC's or what? You getting a link light on your switch or hub? Are you pinging by IP address or hostname. If hostname, make sure they're in /etc/hosts. When you do an ifconfig -a,... are you sure the net interface is up? What do the other devices on the net tell you about the Solaris 8 box? If you can ping your NIC by IP address, and you ping the network broadcast address with ping -s xxx.xxx.xxx.255 (if class C) Do you have a hostname.device_instance# file in the /etc directory.... where device is the exl...whatever and instance # is the instance # of the device? It's probably something pretty easy. Did you try a reboot -- -r after you installed the drivers for the NIC? Good luck!

Peace....

Report this post to a moderator

Can you ping the Solaris box from another host?

__________________
Check out my music at
www.chodan.com
Rural Development in Eastern Ky.
www.centertech.com
"It is our decisions that show us what we truly are in life, not our abilities."

Report this post to a moderator

Perform sysdef -D | grep "elx*" | tee elxfi. The screen output should look like this :

pci10b7,9050, instance #0 (driver name: elxl)
* Loadable Object Path = /platform/i86pc/kernel/drv/elxl
* Loadable Object Path = /kernel/drv/elxl
etc...

grep "elx*" /etc/path_to_inst | tee instancefi . You should see :
"/pci@0,0/pci10b7,9050@11" 0 "elxl"


ls -l /platform/i86pc/kernel/drv/elx* /kernel/drv/elx* > lsfi


Mail me the following file
a. elxfi file in your current directory
b. elxl.conf found in kernel/drv.
c. instfi in your current directory
d. lsfi in your current directory


No promise, see if I could help. Check the earlier suggestion of verifying the file hostname.elx0 etc.

__________________
Either believe what you do or do what you believe.

Report this post to a moderator

I come across it, bingo when I incidentally as I setup my second Solaris (this one on Intel) using 3C595 card, it exhibit the same behaviour.

The masterdb files shows
pci10b7,5950 pci10b7,5950 net pci elx.bef "3Com 3C595 Ethernet 100bTX"

and the sysdef -D output shows
pci10b7,5950, instance #0 (driver name: elx)

there is a corresponding hostname.elx0 and so is an entry pci10b7,5950 in the /etc/path_to_inst. Things relating to default router etc have no relationship as it is from the same network and the ipconfig shows up and running with broadcast. The subnet mask is the same. Nothing to do on networking end.

However the kernel/drv only shows the elxl and elxl.conf and no elx and elx.conf. Even copying the elx and elx.conf from the /platform/i86pc/kernel/drv directory to the kernel/drv and rebooting with reconfiguration does not help. Change the card to 3c509 and it works fine. The card 3c595 is definitely working as when I boot from Windows it works. Believe there is bugs there.

Check your card, if it is 3c595, forget about it and change it.

__________________
Either believe what you do or do what you believe.

Report this post to a moderator