eth2 is UP, but not RUNNING

[Purnelle]

I have a problem with my BL25p running RedHat Enterprise Linux 4 ES U1.

All network card run find,except eth2.
Here the ifconfig of eth0 :

eth2 Lien encap:Ethernet HWaddr 00:13:21:20:AA:A0
inet adr:192.168.0.41 Bcast:192.168.0.255 Masque:255.255.255.0
adr inet6: fe80::213:21ff:fe20:aaa0/64 Scope:Lien
UP BROADCAST MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 lg file transmission:1000
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
Interruption:217 Mémoire:f7ef0000-f7f00000

eth2 not running.
the driver used is the bcm5700 from HP website.

could you explain why this card not work.

[GGA]

hello

in /etc/sysconfig/network-scripts/ifcfg-eth2
is the option
ONBOOT=yes ?

regards gga

[Ivan Ferreira · Accepted solution]

The other network adapters use also the same kernel module?

It's seems a problem with the kernel module:

RUNNING The driver has allocated resources for the interface, and is ready to transmit and receive packets. This is a read-only option that is set by the driver.

Is the network cable working?

What is the output of ethtool eth2? (If it supports ethtool)

[Muthukumar_5]

Check these?

a) netstat -in - Is eth2 ip available here?
b) ifconfig eth2 - what are you getting?
c) ping or nslookup
d) netstat -rn - Gives routing entry
f) /etc/sysconfig/network-scripts/ifcfg-eth2 - what is the content on this?

Post responses on this to pin point problem.

-Muthu

[Purnelle]

Output of ethtool eth2

Settings for eth2:
Supported ports: [ TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised auto-negotiation: Yes
Speed: Unknown! (0)
Duplex: Half
Port: Twisted Pair
PHYAD: 0
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: g
Wake-on: d
Link detected: no

[GGA]

u havent a link on ur card
look at the last entry
gga

[Purnelle]

Muthukumar :
The attached file contain all information what you want

[Purnelle]

Why cmanic send message

Oct 27 14:01:40 mysql01 kernel: bcm5700: eth2 NIC Link is DOWN

and ifconfig says is UP

[Ivan Ferreira]

UP The interface is up. This option is turned on when an address has been configured on the interface.

So, you can get UP but there is no link. Also you can have link but not up, as when you run ifconfig eth2 down.

Verify the network link, ethtool shows that there is no link.

[Purnelle]

I changed the cable and I have the same problem.

When I put the cable in eth2 and in a slot of procurve switch, the corresponding leds lights.

I think that I have a link.

[Ivan Ferreira]

You must be 100% sure that the cable works, test with another computer or laptop. I tell you this because you don't have RX packets, and you should have RX packets always (because of the broadcast) and ethtool reports no link, even when the led is lit.

The time when I had a good link, but not RX packets (and no link as you) was because the kernel module did not worked for my ehternet device, I changed the module and then worked, but that was a 3com adapter. That's why you need to ensure that the cable works, otherway, it may be a problem with the module.

Another time my problem was a bad pin of the network adapter connector, so, the led showed link, but the connection did not worked until we replaced the adapter.

[GGA]

hello
are ur switch configured correctly?
try with a other pc on the same port on the switch.
if ur can ping other systems the configuration of the switch is ok ..
regards gga

[Purnelle]

There are two BL20p and one ML530 G2 connected to this procurve.

[GGA]

i think this is a managed switch there u can disable ports too. i mean that the port where u connect the new server is disabeld or administrative shutdown ...
regards gga

[rick jones]

I'd second all the suggestions about having a good link indication.

An addtional diagnostic tool, similar to ping, but which does not require correctly configured IP is "linkloop" which has been ported from HP-UX to Linux and can be found IIRC on sourceforge.net. You specify a remote MAC address (in the same broadcast domain) and a TEST frame is sent to it. If the remote system is HP-UX it should respond just fine, if not, you may need the linkloop server that comes with the linkloop package.

[Purnelle]

linkloop -i eth2 00:0A:5E:40:00:86
Link connectivity to LAN station: 00:0A:5E:40:00:86 (HW addr 00:0A:5E:40:00:86)
got unexpected packet
PACKET DUMP: data size=99 (0x63)
0000 ff ff ff ff ff ff 08 00 5a 09 05 fb 00 63 e0 e0
0010 03 ff ff 00 60 00 04 00 00 00 00 ff ff ff ff ff
0020 ff 04 52 00 00 00 10 08 00 5a 09 05 fb c2 83 00
0030 02 03 d7 21 4c 45 58 39 30 41 30 44 46 00 00 00
0040 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0050 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0060 00 00 10
END PACKET DUMP

[rick jones]

Well, it at least suggests that something does come-in on the link :) Out of curiousity, was the remote system HP-UX, or something else?

[Purnelle]

It's a ML530G2 with RedHat 8.0pro

but now this bl25p run fine with eth2, I don't know why suddenly eth2 work fine ?

But I have the same problem for my other BL25p

[Purnelle]

From my other bl25p

linkloop -i eth2 00:0A:5E:40:00:86
Link connectivity to LAN station: 00:0A:5E:40:00:86 (HW addr 00:0A:5E:40:00:86)
got unexpected packet
PACKET DUMP: data size=42 (0x2a)
0000 ff ff ff ff ff ff 08 00 5a 09 75 11 00 2a aa aa
0010 03 00 00 00 81 37 ff ff 00 22 00 04 00 00 00 00
0020 ff ff ff ff ff ff 04 52 00 00
END PACKET DUMP

[Purnelle]

I found the problem and the solution.

I forgot that I changed the position of cable for the my first BL. Now, I make the same change with the second Bl and it work fine.

With the BL20P (RH EL 3)
eth3 eth2 ---------------------- eth0 eth1

With the BL25p (RH EL 4)
eth3 eth1 ---------------------- eth0 eth2

[Craig Jackson_3]

Yes, the BL20p G3 and BL25p have different hardware layouts.

The BL20p G3 has 1 ether on the motherboard and 3 on the daughterboard. Because of the way that the PCI bus is arranged, Linux sets up the daughterboard ethers as eth0, eth1, and eth2, and the motherboard as eth3. These are Ethernet 2, 3, 4, and 1, respectively, as enumerated in the ILO. (I.e. the ILO considers the motherboard ethernet to be number 1.)

The BL25p has 2 ethers on the motherboard and 2 on the daughterboard. The way that the PCI is laid out, they're enumerated by Linux in "natural" order. That is, ethers 1-4 as described in the ILO correspond to eth0-eth3 in Linux.

Linux assigns ethN ids in the same order as they appear in the output of "lspci".

Of course, I haven't found a FM to explain any of this.