Lancard stops working ... after a week

[Elmar P. Kolkman]

We have some HP systems with a very heavy network load. They are used as storagenode for a Legato Networker setup, using NFS for storage.
They have 3 1Gb lan cards (actually: 1 on-board and 2 on FibreChannel/Gb-lan combicards) and we have one of those cards failing. We didn't think much about it when it happened a week ago, and did a reboot when we couldn't find the reason, which resolved the issue. But it happened again today...
The system functions fine over the other 2 lans (being the backup and storage lan) but the 3rd (the user lan) doesn't work. Even a linkloop from another system on the same userlan fails for this lan, while the other 2 lan cards respond correctly on a linkloop... So it seems to be caused by something low-level.

We have the ip_ire_gw_probe turned off, so an unaccesseble gateway shouldn't be a possible cause.
unplumb and plumb doesn't solve the issue.
I don't want to reboot the server if it isn't really necessary, but it was to be done if this isn't resolved before 15:00 local time or if someone needs to restore over the userlan (not all clients are connected to backup lan).

Any suggestions anyone?

[Muthukumar_5]

May be physical connecvitity on lan3 will cause problems. IS there any error messages in syslog.log file or daemon.log on legato networker.

Can you check the routing setting and ping connectivity locally to that interface ip-address.

--
Muthu

[Arunvijai_4]

Hello,

Can you check syslog and daemon.log for any discrepancies ? What does the status of the lan3 say ? lanadmin may help you.

-Arun

[Elmar P. Kolkman]

Syslog.log and dmesg don't report anything.
Resetting the port on the network switch didn't help either.
AND: that reset also didn't show up on dmesg either!

Strange thing is I get packets in and out when viewing the lan card with netstat...

Local ping to the card works.

[Arunvijai_4]

Hello,

What do you get when you do a # lanadmin -p ?

-Arun

[Elmar P. Kolkman]

Some other things:
lanadmin shows the card to be running 1000 Mb Full duplex, auto negotiate on (there is now way I know of to set a card to Gb without autonegotiate... Lanadmin doesn't support it.)
Viewing the status with lanadmin doesn't show anything strange.

The card is named lan2 (HP-Ux starts counting at 0).

[Elmar P. Kolkman]

Arunvijai, it reports:
ifconfig
ifconfig
rbootd
rbootd

I stopped the rbootd (without userlan it is not really necessary to be running), and unplumbed the device, after which I ran /sbin/init.d/net start, but the problem persists... I got only two ifconfigs left in the lanadmin -p output, though.

[Muthukumar_5]

what you are getting for,

# ioscan -fnC lan

--
Muthu

[Elmar P. Kolkman]

Muthu,

It reports all cards as being claimed:
root> ioscan -fnC lan
Class I H/W Path Driver S/W State H/W Type Description
========================================================================
lan 0 0/1/2/0 igelan CLAIMED INTERFACE HP PCI 1000Base-T Core
lan 1 0/3/1/0/6/0 igelan CLAIMED INTERFACE HP A9784-60001 PCI/PCI-X 1000Base-T FC/GigE Combo Adapter
lan 2 0/4/1/0/6/0 igelan CLAIMED INTERFACE HP A9784-60001 PCI/PCI-X 1000Base-T FC/GigE Combo Adapter

[Arunvijai_4]

May be this thread could help you,

http://forums1.itrc.hp.com/service/forums/bizsupport/questionanswer.do?threadId=113391
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=9001

-Arun