lan0 net interface state down

[radhat]

I'm new to HP UX so forgive me if I'm rehashing. I'm tried to do my homework here and have tried many of the excellent solutions in other posts. However, I'm stuck.
I believe the HP version I'm using is 10.2
but I'm not sure because I don't know how to verify it.
I'm trying to connect to my internal network on a subnet of 255.255.255.0 using 192.168.1.x

IP address of the HP UX Box is 192.168.1.2
LanScan shows the following
Hardware Path 2/0/2
station address: 0x0800097bXXXXX (my X's)
Crd in# 0
Hardware State UP
Net-Interface
Name Unit lan0
State DOWN (I think this may be my problem)
NM ID 4
Mac Type Ether
HP DCPI Support YES

Netstat -in shows:

Name MTU Network Address
ni0* 0 none none
ni1* 0 none none
lo0 4608 127 127.0.0.1
lan0 1500 192.168.1 192.168.1.2

IFCONFIG shows
lan0 flags=862
inet 192.168.1.2
netmask ffffff00
Broadcast 192.168.1.0

etc/rc.config.d/netconf shows:
interface name=lan0
ip address= 192.168.1.2
subnet=255.255.255.0
broadcast=192.168.1.255
Lanconfig_args=ether
dhcp_enable=0

route destination=0.0.0.0
route mask=""
route gateway=192.168.1.1
route count=1
route arg=0
gated=0
gated_args=0

Physcical connection: RJ45 cable (tested myself with tester. Works)
I am using a UNICOM Mini 10 Base T Transceiver to convert from AUI to 10 Base T and is set SQE switch is set to on (as per manual).
Cable length is approx. 3 feet.

I don't know what else to tell you.

Thanks
RAD

[radhat]

forgot to tell you...

Can't ping anything other than 192.168.1.2
Get a No route to host when I try

[kamal_9]

Hi RAD
just try
#ifconfig lan0 down
#ifconfig lan0 up

[radhat]

5 Points for the fast answer (and technically correct).
Your answer did result in a change of interface status, however, I still get "No route to host" when trying to ping even internally.

Any ideas on that?

Thanks

RAD

[Steven E. Protter]

You have a problem with your network configuration design.

The problem here is that the ip address and network mask being assigned are not fully compatible with your network.

Another hpux-ism to know. Two lan cards can not be brought up on the same network/subnet.

This will bring down or paritally disable networking. This means there can not be two cards up on the 192.168.0 network (mask 255.255.255.0)

This is something you CAN get away with on Linux and Microsoft servers.

There is a flaw in your network design and you need to post more info or see your network administrator to increase your knowledge of setup.

I agree this is not cabling.

You should also try bringing up the card with ifconfig. Once you get it right you can change your /etc/rc.config.d/netconf files.

SEP

[radhat]

only one card. I went back and doublechecked.
Are you seeing something that leads you to believe there are 2 cards? Maybe I'm missing something.

[Ron Kinner]

If I remember correctly
uname -a
will tell you what version you are running.

What does

lanadmin
lan
display

show? Do you see two UPs in the first few line? Two UPs means the NIC is operational from the software point of view and Ethernet connection is good. One Up and one Down usually means the Ethernet connection is bad. So you would look at the cable or the converter or the hub/switch or the NIC. Two Downs means the NIC did not get programmed correctly or is in a down state for some other reason.

Do you have a hub or a switch? Or are you trying to get two machines to talk directly to each other. In the last case you must have a crossover cable.

If you are running a switch make sure it is either set to auto or to 10 half.

IF the NIC also has a BNC coax connector on it there may be a switch on the NIC which selects between the AUI and the BNC. You may have to remove the NIC to get to it.

Ron

[radhat]

Ron you have a point.
Going back to look shows. interface status as down again. I bring it up and within 2 minutes it's down again.

Starting to sound like hardware to me.
I was going into a 10/100 hub switched over to the router itself, no change.

Version is B.10.20 A 9000/715(thanks for the tip)

What Nic can I put in this box (never cracked an HP case before)

[Mic V.]

Have you verified the settings on whatever network gear is connected to the other end of your cable (presence of link, correct subnet/VLAN, duplex, speed, ...) and that the transceiver is still working? My gut feeling is that it's on that side, not yours.

As a side note, not directly related to this behavior -- I'm assuming you approximated the contents of netconf rather than pasting it. :) The text in your original quote shouldn't work.

HTH,
Mic

[Mic V.]

Sorry, forgot your last question. If I'm interpreting uname correctly that this is an HP 715 workstation, all I can say for sure is that there was a FDDI card for it at one time (I had one). Unfortunately, my usual way to find out what's available for old hardware is to ask a sales rep whose been around long enough. I'm not sure where people would go to find the info (you can find such info about currently-sold models easier).

This will help find information about NICs if you have a P/N:

http://partsurfer.hp.com/

Mic

[radhat]

Upon further examination I have link light on the transciever, and on both the hub and the router (whichever I plug into).
Router is a Lynksys 10/100, auto negotiate.
Hub is 10/100 Auto negotiate.

Hmm.... I wonder...

Could this be a crossover cable issue?

Please sanity check me on this and tell me if I'm making sense.

Assuming that this is not a Nic hardware issue, and that the ports on the hub and router are truly auto, then a cross over cable would only confuse the issue...

[Jeroen Peereboom]

I think you have a lot of hints now.
If you have a link led on, the cable is correct (not cross).

Why don't you try to setup a simple network to rule out difficult network issues.
- Your HP-UX workstation
- A PC
- A hub between them (10Mbps is good enough)

Configure the PC and the workstation on the same subnet (192.168.1.0).
Try to ping and telnet to the workstation.

JP


JP

[radhat]

Nope doesn't work. Have link lights still no connection.

C:\DOCUME~1\RADHAT>ping 192.168.1.2

Pinging 192.168.1.2 with 32 bytes of data:

Request timed out.
Request timed out.
Request timed out.
Request timed out.

Ping statistics for 192.168.1.2:
Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

C:\DOCUME~1\RADHAT>telnet 192.168.1.2
Connecting To 192.168.1.2...Could not open connection to the host, on port 23.
A connection attempt failed because the connected party did not properly respond
after a period of time, or established connection failed because connected host
has failed to respond.

[sinhass]

See though I'm not also an expert but try to change all the setting thru #set_parms initial and recheck.
-sinhass

[Ron Kinner]

There is a nettl command

/usr/sbin/nettl -start
/usr/sbin/nettl -log i -e all

which should create a log file:

/var/adm/nettl.LOG00

which may be able to tell you why the NIC is going down.

I think this is a replacement transceiver:

28641-60015
HP ThinLAN (10Base-2) transceiver - Has AUI and BNC connectors - (28641B) - Plugs into the AUI port

It's $66 from HP but I'm sure you can find used ones. They no longer offer any network cards.

I think it fits on the 28640a which should be an EISA HP-PB 10T which should be a replacement for the NIC but I can't swear to it. I can ask our HP users tomorrow. They may even have a few 715s lying around.



Ron

[Sridhar Bhaskarla]

HI Rad,

From DTE to/from DCE - straight
From DTE to/from DTE or DCE to/from DCE - crossover.

In your case your workstation (DTE) is going to the hub (DCE). So your cable is fine. Can you set the SQE off and see if you see any improvement?. In ethernet1 environment, SQE test signal is treated as an error.

Also, while you are resetting the interface using 'landiag', do you see the link lights go down and up on the tranceiver?.

-Sri

[Fabio Ettore]

Hi,

can you explain what is

route destination=0.0.0.0

in netconf?

Try

# route add default 192.168.1.1 1

After this command can your system reach other networks?
If yes then change

route destination=0.0.0.0
to
route destination=default

Best regards,
Ettore

[radhat]

Guess it's time for an update.
Thank you all for the time you've dedicated.

By following everyone's suggestions, we have proceeded past the "no route to host message". So I will be assigning 5 points to everyone that I haven't submitted points for yet.

However, the problem is still not fixed.
Now when I ping I just get a flashing cursor.
When I break it still shows 100% packet loss.

Ron K. using the nettl command results in a message about "failed to start console logging due to invalid console logging option file /var/adm/console.opts. Ichecked the files this message refed to. The log said to refer to line 3 of console.opts, because the date is wrong. Line 3 of of console.opts shows at date template (looks like) the final bit is puzzling it shows 01/30/103. Message in error log states it should be between 0-99 (I figure this as a Win2k bug).

When trying to open /var/adm/nettl.LOG00 seems to hand for a good amount of time, and when it returns, the output is not something that I'm able to read.

Sridhar: Lights on both the Hub and the Transciever, very little blinking. The hub also shows it as 100 Mbps.
Switching from SQE on to SQE off didn't seem to do a thing.

Ettore: change route destination to default in netconfig and did a net stop and net start. No change.

Some of the things I noted while working through your suggestions.

The CDE (common desktop enviorment) seems to not to pleased with the changes we are making. It seems to hang from time to time especially when trying to force something over the wire.
Often times I can't log in unless I disconnect the RJ45 cable.

rc.log is showing lan0 as not fount and I also saw a line stating that localhost name is not qualified.
I noted that there is output from /sbin/rc1.d
/sbin/rc2.d and /sbin/rc3.d
All have identical output:
for example; output from /sbin/rc1.d
s320 hostname[36]: lan0 not found
s320 hostname[36]: 192.168.1.2 not found
s320 hostname[36]: half not found
Wed, Jan 28 22:49

Ideas? I'm plum out. :)

Thanks again.

Ron

[Chris Watkins_1]

Did you ever run the command Ron asked about?

run:
# lanadmin
Enter command: lan
Enter command: display


And show us what you have there.
The light that shows you "100" must be incorrect... is my thought.

Afterthought... is the hub capable of running 10 AND 100 connections?
Some will run them simultaneously... some won't.
In that case, you'd have to plug in this workstation first,
then everything else. (tying it all to 10mbit)

Either that, or isolate this machine by plugging it into
its own hub... then connecting from that hub to the.(or a switch)

It may or may not work... it solved a similar problem with my
712/80 workstation at home, using 3Com hubs, however.

[radhat]

Chris: Hub can do both at the same time.
I have some machines running at 10 others running at 100.
I have tried seperate ports with no change.

As requested. LAN ADMIN

Lan Interface display
Thu Jan 30, 2003 13:43:28

Network Management ID= 4
Description= lan0 Hewlett-Packard Lan Interface Hw rev 0
Type(value)-ethernet-csmacd(6)
MTU Size=1500
Speed=10000000
Station Address=ox80097b9dd6
Administration Status(value)=UP(1)
Operational Status(value)=UP(1)
Last change= 4301820
Inbound Octets 3933000
Inbound Unicast Packets = 0
Inbound Non-Unicast Packets=320
Inbound Discards=0
Inbound Errors=0
Inbound Unknown Protocols=0

Outbound Octests=5914224
Outbound Unicast Packets =0
Outbound Non-Unicast Packets=320
Outbound Discards=11
Outbound Errors=0
Outbound Queue Length=0
Specific=655367

Ethernet-like Statistics Group
Index=4
Alignment Errors=0
FCS Error=0
Single Collision Frames=0
Multiple Collision Frames=0
Defered Transmissions=0
Late Collisions=0
Excessive Collisions=0
Internal MAC Transmit Errors=0
Carrier Sense Errors=8623
Frames to long=0
Internal MAC Receive Errors=0

(transcribed long hand. Data is correct even if Puncuation and spelling are not).

:)

[Chris Watkins_1]

Hate you did that in longhand... checkout the "script" command:

# script some_output_file

Starts a script session, with output name that you specified.
Once you're finished with the command, you can then
hit CTRL-D to end it, and edit the output_file for cleaning.
Then it's a simple matter of attaching or copy-paste :)

Anyhow... maybe we need to see your actual netconf file.
This is beginning to be a puzzle... but given the errors in
your rc.log, I'm wondering if something in the netconf is hosed.
Also, make sure you don't have any old netconf files
laying around in /etc/rc.config.d/. If you want to save a
copy before editing, it needs to go somewhere else.

Here's an example of a (cleaned up) netconf for your machine:
(Note that I've used "0xffffff00" and "255.255.255.0"
interchangeably, and never seen any difference, for sub-mask. Not sure on 10.2)

########################################

HOSTNAME="your_hostname_goes_here"
OPERATING_SYSTEM=HP-UX
LOOPBACK_ADDRESS=127.0.0.1

INTERFACE_NAME[0]="lan0"
IP_ADDRESS[0]="192.168.1.2"
SUBNET_MASK[0]="255.255.255.0"
BROADCAST_ADDRESS[0]="192.168.1.255"
INTERFACE_STATE[0]=""
DHCP_ENABLE[0]=0

ROUTE_DESTINATION[0]="default"
ROUTE_MASK[0]=""
ROUTE_GATEWAY[0]="1"
ROUTE_ARGS[0]=""

GATED=0
GATED_ARGS=""

RDPD=0

RARP=0
#########################################


Maybe just compare yours, to see if there's any serious discrepancy.
(blanks, missing underscores, etc...)
Might save you some typing, if yours looks a lot like this,
after you ignore all the comments. Let us know.

[Ron Kinner · Accepted solution]

Your 103 is a Y2K problem as you guessed. You can probably get a patch for your system to fix it. I'm at home now so don't have my references. Check the man on nettl to see if there is a program that formats the log file for you.

The lanadmin command shows:

Carrier Sense Errors=8623


This implies a layer 1 problem. (Cable/connector is bad. Tranceiver is bad. Hub port is bad or possible the NIC is bad.) The Carrier Sense is what tells the NIC that the circuit is live. Either the signal is not there or the detector is faulty. I used to have a Compaq server that had a similar problem. The RJ45 was bad and the link light would go out when you wiggled it. We fixed it by duct taping it so that the cable couldn't move until we got a new NIC. You might try a new tranceiver or flip the SQE switch on the transceiver to the other position and see what happens.

Ron

[radhat]

Ron Got it!!

Bad Transciever.

Replaced it worked.

Wasn't the cable. I retested with a different cable and it still works.

Wasn't the router or hub, I tried it in both.
Both work.

It's a very good possiblility that it might have been both a hardware issue (the transciever) and software.

The only other change I made was with the monitor (I switched over to the HP Monitor (A4033A). At least the colors look good. It was just this nasty green with my NEC.

Now to figure out DNS, patches (where are they BTW, and can I ftp there from the HP box?) Can I upgrade to 11 for free or is it a paid upgrade. Do I really really want to?

Seems now the work begins.

Thank you ALL for the help you provided. Couldn't have done it without you. :)

Ron

[Mic V.]

All patches can be gotten through:

http://www1.itrc.hp.com/service/patch/mainPage.do?admit=-682735245+1075608469424+28353475

I suggest you first get the latest 10.20 patch bundle and work from there. I'm still trying to find whether they're online (too old?). They're not in the "current" area or the "old" area.

The upgrade to 11.11/11i is free. It gets you more dynamic kernel parameters and other things. Since you're running 10.20, I suspect you're on a server that's too old to take advantage of some of the neat new stuff (partitioning).

Mic

[Mic V.]

Addendum: Sorry, yes, you can ftp:

ftp.itrc.hp.com

Mic