Newbie questions

[Frederic Wartner]

Hi all,
I'm trying to connect a 715/50 10.20 via Ethernet to a Linux PC, which I know is OK because everything works fine with another PC.
I specified 192.168.0.4 in "set_parms ip_address", but SAM shows the address 0.0.0.0 for lan0, and when I modify it and enter 192.168.0.4, it says the address already exists and this can cause problems... So I'm confused: is there an IP address for the machine and another one for the network card?
Right now the interface is set back to 0.0.0.0. When I boot with the network cable connected, the sequence is very long and just before CDE starts, the system reboots, and so on. With the cable disconnected, it boots normally and CDE starts. In both cases, the DNS client service fails during boot sequence.
Would you be kind enough to help me, or to let me know where I can find these informations on the Web?
Thanks in advance.
Best regards
Frederic

[Steve Steel]

Hi

Look in

/etc/rc.config.d/netconf

This is what defines your networking

Steve Steel

[Michael Duthie]

Frederic,

Is 192.168.0 not the subnet MS use for ICS ? if this is what your trying to set up the WKS should be setup for DHCP.

If not might be useful to give a print of

lanscan

more /etc/rc.config.d/netconf

ifconfig lan0

[melvyn burnard]

sounds to me like you are using a subnet that does not allow this address.
post the output of netstat -in, ifconfig lan0, and your netconf file.

[Michael Steele_2]

Gather the following:

Host ip
Host name
Netmask (* This may be your problem *)
Gateway ip
DNS ip (* if exists *)
NIS server ip (* if exists *)
NTP server ip (* if exists *)

set_parms ip_address

enter all but hostname

set_parms hostname

verify and set if needed.

reboot

ping your hostname/ip
if ok
ping your gateway
if fail
check cable to switch or hub
ping gateway again

[Steven E. Protter]

make sure the network card is claimed.

ioscan -fnC btlan

or

ioscan -fnC


You have link light on the card and the switch?

You should get an IP address either way.

SEP

[Ron Kinner]

Looks to me like you may be trying to do DHCP without realizing it and without a DHCP server.

You could also have an old version of SAM which needs a patch.

to find out what you are really using for an IP address do

netstat -anv

Ron

[Paul Sperry]

Everything is in /etc/rc.config.d/netconf

but also take a look at your

/etc/hosts
/etc/resolv.conf
and
/etc/nsswitch.conf

files as well

[Frederic Wartner]

Here's my netconf:

HOSTNAME="hpux"
OPERATING_SYSTEM=HP-UX
LOOPBACK_ADDRESS=127.0.0.1
INTERFACE_NAME[0]=lan0
IP_ADDRESS[0]=192.168.0.4
SUBNET_MASK[0]=255.255.255.0
BROADCAST_ADDRESS[0]=
LANCONFIG_ARGS[0]=ether
DHCP_ENABLE[0]=0
ROUTE_DESTINATION[0]=default
ROUTE_MASK[0]=""
ROUTE_GATEWAY[0]=192.168.0.3
ROUTE_COUNT[0]=1
ROUTE_ARGS[0]=""
GATED=0
GATED_ARGS=""
RDPD=0
RARPD=0

Thanks to all for your help.

[Frederic Wartner]

Also lanscan:

Hardware path 2/0/2
Station address 0x080009786362
Crd In# 0
Hardware state UP
Net-Interface Nameunit lan0
State UP
NM ID 4
MAC type ETHER
HP DLPI support Yes
Mjr num 52

and ifconfig lan:

lan0: flags=863
inet 192.168.0.4
netmask ffffff00
broadcast 192.168.0.255

[Michael Duthie]

all looks ok.

is the hostname in /etc/hosts with the correct IP ?

can you ping yourself?

can you ping the loopback addr?

can you ping the broadcast addrs?

does nslookup work ?

[Michael Steele_2]

Basic rule of being a system administrator - DON'T TROUBLE SHOOT WITH SAM!!!

Sound like everythings there but some SAM patches.

You can ping your gateway, right?

Use these commands for trouble shooting:

lanadmin -x 4 (* Current speed and duplex mode must match with switch *)

lanadmin -X 100FD 4 (* Set to Full Duplex 100BaseT *)

netfmt -LN -f /var/adm/nettl.LOG* > /tmp/file

vi /tmp/file and search for disasters and errors

Match and verify the next two:

what /stand/vmunix | grep -i btlan

swlist -l product | grep -i 100

[Frederic Wartner]

No, I can't ping the other machine.
When the system boots with the LAN cable connected, the DNS client service fails and it reboots again and again so I don't have a chance to try pinging. Meanwhile, I can ping it from the Linux box by IP and by name .
When it boots with the LAN cable disconnected, it pings itself by IP and name.

[Frederic Wartner]

I apologize: this is not the DNS client service that fails during the boot sequence, it's the NFS CLIENT SUBSYSTEM...
Sorry for this mistake.

[Michael Steele_2]

A) Check your cable connection to the switch.

B) Obtain the MAC address of the switch port.

C) Verify with arp -a

D) linkloop MAC (* of switch *)

F) Obtain the speed and duplex of the switch

E) Set your LAN to same with: lanadmin -X 100FD 4, or, lanadmin -X 10FD 4, or, lanadmin -X 100HD 4, etc.

NOTE: Depending on the switch there may be some latency before connectivity. Maybe an hour. Check with the manufacturer or network admin.

G) Ping the default gateway router

[Michael Steele_2]

Note: Its important for D) linkloop MAC (* of switch *)...to work. If failure then change cable or port on switch.

[Frederic Wartner]

Well, I'm afraid there's no switch. The PC card and Wks Ethertwist transceiver are connected by a crossed cable (I'm not sure this is the correct word in English), which works fine between 2 PCs.
Also I'm not sure I explained clearly that the NFS client subsystem starts without the cable, and doesn't with it.

[Michael Duthie · Accepted solution]

try disabling DNS, by renaming /etc/resolv.conf

and disable NFS by editing /etc/rc.config.d/nfsconf and changing the 1 ofr a 0 at NFS_server, NFS_client and Start_mountd.

that should fix the rebooting problem.

[Michael Steele_2]

OK, cross over cable, now I get it. Peer to peer network without a gateway. No default gateway needs to be specified.

Disable every network utility you don't need including: /etc/rc.config.d/named, and the default route in netconf.

[Frederic Wartner]

It's OK after removing all useless utilities.

Thanks to all who helped!

[Michael Duthie]

DNS, NFS Useless ? How dare you ! Your a "Newbie" alright. lol

Glad you got it working.