Problem with Network when system board replaced

[Nipun Patel]

Hi,
Recently our system board for DS25 had to be replaced.

We were using EIA0 (Fast/Ethernet Card) for connection to the internet or TCP/IP. Now with the new board (we were unable to save the SRM console settings from the previous board) we cannot connect to the internet. Somehow it seems the eia0-mode settings do not seem to match or there is a problem with the port

We havee static IP and have the following options for the mode

Auto-Sensing
Twisted-Pair
Full Duplex, Twisted-Pair
AUI
BNC
Fast
FastFD (Full Duplex)
Auto-Negotiate


I have tried with
1) Auto Negotiate
2) Full Duplex, Twisted Pair

For 1) I get
***Error (eia 0.0.0.8.0) No link,Auto Negotiation did not complete The
selected mode is illegal or incompatible.

For 2) I do not get any other message but when I boot up sometimes the internet connection seems fine and other time it hangs. This suggests that the settings are in correct.

Would setting a different speed and testing for each mode one at a time be the best option ?

As far as the switch is concerned the network people informed me that they are set at 100/Full

[Ian Miller.]

If the switch is fixed at 100/Full then set the card to FastFD

[Robert Brooks_1]

Unless there is a problem with the switch, any modern version of VMS should correctly autonegotiate the speed and duplex settings.

This comes directly from the VMS engineer who works on the ethernet drivers.

-- Rob

[Peter Zeiszler]

Ian had it closest. Actually contact your network admin and have they force the port for your system to 100/full then force your card to 100/full.

If you have the equipment you can also have the network guys test your network speed and duplex at the cable where it plugs into your system.

Check your speed and duplex at the boot time and after it has booted. I like using "mc lancp show dev ewa0 /param".

We have ran into multiple instances where auto negotiate does not work and also where forcing the card to a specific speed without forcing the switch didn't get them working together.

[Andy Bustamante]

Peter is correct, the settings on the Alphaserver and the switch need to match. You can change the Alphaserver with the LANCP utility and avoid rebooting.

$ MCR LANCP
LANCP> SET DEVICE EIA0/SPEED=100/FULL_DUPLEX

To make the changes permanent, replace SET with DEFINE.

You can use SHOW DEVICE/COUNTERS in LANCP to check for errors, this can be a sign of a duplex mismatch.

Andy

[Steven Schweda]

My experience is limited to stupid switches,
but I've never had a problem with
Auto-Negotiate.

Are you sure that nothing bad happened to the
E-net card or the cable (or the switch) when
the work was done to the system?

[Eric Dittman]

If you hard-code the settings on one end of the connection you need to hard-code the settings on the other end. Autonegotiation doesn't work right if you have one end hard-coded and the other set to autonegotiate.

[nipun_2]

Ian, Peter and Andy
you guys are geniuses....

Doing so worked perfectly. Also the idea mentioned by Peter to do it in the terminal is very helpful. I was always shutting down the node.

So now I have got 2/3 nodes to work...and got all the switches set to 100 /FullFD

One node is not working so I got the following mail from my network admin
00-02-a5-20-07-5b - shows connected but is not visible on the router and is not pingable... It occasionally shows as an incomplete arp... Check it has gateway xx.xx.xxx.x and subnet mask yyy.yyy.yyy.yyy...

So I guess I can do this from lancp or tcpip?
I will try to figure out how to check those settings and if required modify them..so that the last node works. I believe this would be now be openvms issue and not the hardware issue as all of them are on FastFD

Nipun

[Volker Halle]

Nipun,

the first check would be $ TCPIP SHOW INT
Make sure, there is a valid interface defined with the expected address and network mask. The packets received and send counters should be non-zero.

Next you could check the LAN device counters with LANCP> SHOW DEV/COUNT EIA0 - note that you can also look at the LAN driver console messages with LANCP> SHOW DEV/INT EIA0 (will display internal counters and the console messages at the bottom).

Volker.

[nipun_2]

Hi Volker,
I executed the commands you mentioned and it seems to indicate that I don't have EIA0 configured as I do not even get that name as can be seen in the log below (Also attached as Text File)

Note: I have two ethernet cards WE0 is already congfigured for cluster. I need to use the EIA0 for connecting to the internet.

So I believe I need to figure out the correct options with

$@tcpip$config


========Log of the current network settings ============ (Also attached as Text File)

$ tcpip show int
Packets
Interface IP_Addr Network mask Receive Send MTU

LO0 127.0.0.1 255.0.0.0 319 319 4096
WE0 10.35.221.122 255.255.255.128 2543 2563 1500



LANCP> show dev /count EIA0

Device Counters EIA0:
Value Counter
----- -------
254989 Seconds since last zeroed
20116718763 Bytes received
9590441257 Bytes sent
62529329 Packets received
36359847 Packets sent
31946897 Multicast bytes received
16878723 Multicast bytes sent
228300 Multicast packets received
141932 Multicast packets sent
0 Unrecognized unicast destination packets
6587 Unrecognized multicast destination packets
0 Unavailable station buffers
0 Unavailable user buffers
0 Alignment errors
0 Frame check errors
0 Frame size errors
0 Frame status errors
0 Frame length errors
0 Frame too long errors
0 Data overruns
0 Send data length errors
0 Receive data length errors
0 Transmit underrun errors
0 Transmit failures
0 Carrier check failures
0 Station failures
0 Initially deferred packets sent
0 Single collision packets sent
0 Multiple collision packets sent
0 Excessive collisions
0 Late collisions
0 Collision detect check failures
LANCP>