Operating System - HP-UX
1849239 Members
2930 Online
104042 Solutions
New Discussion

Re: Serviceguard A.11.18.00 lan problem

 
SOLVED
Go to solution
wkg
Occasional Advisor

Serviceguard A.11.18.00 lan problem

Hi all,
I have 4xRX6600 IA64 identical servers.

Below my hardware LAN:
ioscan -fnC lan
Class I H/W Path Driver S/W State H/W Type Description
===================================================================
lan 0 0/2/1/0 iether CLAIMED INTERFACE HP A7012-60601 PCI/PCI-X 1000Base-T Dual-port Adapter
lan 1 0/2/1/1 iether CLAIMED INTERFACE HP A7012-60601 PCI/PCI-X 1000Base-T Dual-port Adapter
lan 2 0/4/2/0 iether CLAIMED INTERFACE HP AB352-60003 PCI/PCI-X 1000Base-T Dual-port Core
lan 3 0/4/2/1 iether CLAIMED INTERFACE HP AB352-60003 PCI/PCI-X 1000Base-T Dual-port Core

/etc/cmcluster# netstat -ni
Name Mtu Network Address Ipkts Ierrs Opkts Oerrs Coll
lan3 1500 148.1.0.0 148.1.100.71 302977 0 145274 0 0
lan1* 1500 none none 0 0 0 0 0
lan0 1500 10.70.1.0 10.70.1.71 239588 0 126263 0 0
lo0 32808 127.0.0.0 127.0.0.1 5482 0 5482 0 0
lan3:1 1500 148.1.0.0 148.1.100.81 3050 0 0 0 0

lan0 is heartbeat interface and connected
to separated VLAN.

My SG lan section:

NODE_NAME sprod1
NETWORK_INTERFACE lan3
HEARTBEAT_IP 148.1.100.71
NETWORK_INTERFACE lan0
HEARTBEAT_IP 10.70.1.71
NETWORK_INTERFACE lan1

ifconfig lan1
lan1: flags=842
inet 0.0.0.0 netmask 0


when I run cmcheckconf I receive error:

Begin cluster verification...
Checking cluster file: ZMCSG_RX.ascii
Checking nodes ... Done
Checking existing configuration ... Done
Gathering storage information
Gathering network information
Beginning network probing (this may take a while)
Completed network probing
lan1 on node sprod1 cannot be configured in the cluster
because it does not have an IP address, and it is not a standby lan for any other lan.
Failed to evaluate network

It's not wire problem as soon as I added manually IPs to lan1 on all hosts communication
is OK

Whats wrong ???
13 REPLIES 13
Mark McDonald_2
Trusted Contributor

Re: Serviceguard A.11.18.00 lan problem

Have you tried an unplumb lan1 before cmchkconf
wkg
Occasional Advisor

Re: Serviceguard A.11.18.00 lan problem

Yes, but the same effect ;-/
Mark McDonald_2
Trusted Contributor

Re: Serviceguard A.11.18.00 lan problem

Its a long time since I built a SG system

Can you try changing the order of the networks in the ascii file like so:

NODE_NAME sprod1
NETWORK_INTERFACE lan0
HEARTBEAT_IP 10.70.1.71
NETWORK_INTERFACE lan3
HEARTBEAT_IP 148.1.100.71
NETWORK_INTERFACE lan1
wkg
Occasional Advisor

Re: Serviceguard A.11.18.00 lan problem

Yes I done this but error the same.
Solution

Re: Serviceguard A.11.18.00 lan problem

Being able to put an IP on an interface and ping it procvs nothing beyond physical connectivity.

So you have lan1 and lan3 on the same VLAN yes?

First off you might want to check link level connectivity between lan 1 and lan3.

Get the MAC address for lan3 from lanscan and then try a link level connection using:

linkloop -i 1

I suspect that won't worth either though.

The other point to consider is "how is your VLAN tagging done in your network?" Most network admins tag based on the port on the switch (i.e. if you plug into port 4 you're in VLAN 2 or whatever), but it is also possible to tag packets on a specific VLAN based on what protocol they use (IP, IPX LLC etc.), and even what source IP subnet they use. If you use either of these 2 mechanisms in your network, you'll realise that link-level packets that come from an interface with no IP address are not going to get tagged as being on the correct VLAN and won't find thir way to the other ports in the VLAN.

If this is the case you have 2 choices:

1) Talk to your network team and get them to give you a port based VLAN for lan1 and lan3 on all 4 servers.

2) Configure VLAN tagging on the hosts by creating a virtual interface which tags packets into the correct VLAN before they leave the host. You'll need to know the numeric ID of the VLAN from your network admin and then use either nwmgr or lanadmin (depending on whether you are on 11iv3 or 11iv2) to setup the VLAN. Manuals are here:

http://docs.hp.com/en/netcom.html#Virtual%20LAN

... or of course you could have some other sort of network problem entirely - but this is my guess for your problem.

HTH

Duncan

I am an HPE Employee
Accept or Kudo
wkg
Occasional Advisor

Re: Serviceguard A.11.18.00 lan problem

Hi Duncan,

Thank you for replay.

Before send post I tested linkloop an get
error:

linkloop -i 1 0x001A4B088EDB
Link connectivity to LAN station: 0x001A4B088EDB
error: get_msg2 getmsg failed, errno = 4
-- FAILED
frames sent : 1
frames received correctly : 0
reads that timed out : 1

but my network team ignored this and connect to lan1 service laptop network links works OK

As soon as I will check next two points

wkg

Re: Serviceguard A.11.18.00 lan problem


>> but my network team ignored this and connect to lan1 service laptop network links works OK

And presumably that service laptop had an IP address in the 148.1.0.0 subnet? That of course would work if you're doing protocol or subnet based VLAN tagging.

Another point to check is that your NIC is negotiating to a suitable speed with the switch... check using "landamin -x 1"

HTH

Duncan

I am an HPE Employee
Accept or Kudo
wkg
Occasional Advisor

Re: Serviceguard A.11.18.00 lan problem

Hi Duncan,

Yes service laptop had IP with lan 148.1.0.0
and my network team checked configuration and
says that is OK but I'm not very sure obout ..

Below out with lanadmin:

lanadmin -x 1
Speed = 1000 Full-Duplex.
Autonegotiation = On.
wkg
Occasional Advisor

Re: Serviceguard A.11.18.00 lan problem

Hi again,

I tested linkloop to other RX2600 server:


sprod1: /etc/cmcluster# linkloop -i 3 0x001CC4FBEFB3
Link connectivity to LAN station: 0x001CC4FBEFB3
-- OK
sprod1: /etc/cmcluster# linkloop -i 1 0x001CC4FBEFB3
Link connectivity to LAN station: 0x001CC4FBEFB3
error: get_msg2 getmsg failed, errno = 4
-- FAILED
frames sent : 1
frames received correctly : 0
reads that timed out : 1
sprod1: /etc/cmcluster# linkloop -i 0 0x001CC4FBEFB3
Link connectivity to LAN station: 0x001CC4FBEFB3
-- OK

Re: Serviceguard A.11.18.00 lan problem

Which shows that those interfaces with an IP address on are able to link to another host... again I susepct this s down to how your VLAN is configured in your network. Ask your network admin what "type" of VLAN membership they have configured - the following might help you understand the different types:

http://nislab.bu.edu/nislab/education/sc441/six/VLAN%20types.htm

HTH

Duncan

I am an HPE Employee
Accept or Kudo
Mark McDonald_2
Trusted Contributor

Re: Serviceguard A.11.18.00 lan problem

wkg

To prove to the network guys, if the machines are in the same room, connect the lan ports to a stand alone hub or switch with no VLAN config.
wkg
Occasional Advisor

Re: Serviceguard A.11.18.00 lan problem

Hi Duncan,

One minute ago my network team added all
cards to the same vlan and cmcheckconf
looks OK. You are rigt that VLAN problem
but my's lan admin don't konow what's wrong
but is yours problem.

Thank you for all responses, the problem is solved.

HTH

wkg
wkg
Occasional Advisor

Re: Serviceguard A.11.18.00 lan problem

All OK read above

wkg