cluster configuration file!!

[astra]

Hello All,

i would like to know that how serviceguard "cmquerycl" command identifies the Network interfaces and other HW devices on the server and create a cluster configuration. The reason for asking this question is that we have two servers with the same hardware and software version but the "cmquerycl" command generates different cluster configuration file.

For example: -

On one setup

# Warning: There are no standby network interfaces for lan0,lan3
# Warning: There are no standby network interfaces for lan1,lan2.

Another setup: -

# Warning: There are no standby network interfaces for lan0.
# Warning: There are no standby network interfaces for lan1.
# Warning: There are no standby network interfaces for lan2.

Please clarify ??

Regards,
Amit

[sujit kumar singh]

Hi Amit,


the cmquerycl command checks by using lot many commands with remsh and Superuser authority on both the nodes to gather and have a look of configuration and is intellligent enough in suggesting after analysis according to the best possible High availability config(Obviusoly on the basis of the Hardware limitation)

if you have identical server then check out that you have the Hearbeat Bridged network and that also the LAN Bridged Network in your LAN properly configured.

Then the cmquerycl shall auomatically give you the best possible Standy LAN Configs.


please check using lanascan and netstat and linkloop if possible both sam/smh that.

seems that on setup LAN setup needs o be checked by the Hardware LAN Specialists.

It simply means that lan3 on second setup physically has been connected to the network switches and all in such a way that it might have a standby LAN I/f on second setup but not on the forrst system.


yuse linkloop to check the Physical connectivity in the pair of the Standby LANS.
regards
sujit

[sujit kumar singh]

Hi


If there is sucess in creation of the Cluster ASCII File please post the contents of that file as well as the screen messages that you get there when you invoke the cmquerycl command.


regards
sujit

[Ganesan R]

Hi,

cmquerycl command will query the specified nodes to get the network information and LVM informations.

It uses linkloop to find the bridged network and based on that it will decide the standby interfaces and create the ascii file. There you can define which interface/ip you want to use for stationary and heartbeat ip's.

Basically standby interfaces should be connected to same bridged network(assume same switch)

[astra]

Hello Sujit,

There is not any problem in file generation, rather i wanted to know that why on some setup it writes in file like "lan0" only
and on some setup it writes like "lan0,lan3". Why there is warning on separate lines for each interfaces, whereas on some setup, in the same line for lan0,lan3 and lan1,lan2.

Whats the sequence of interface information generation??

Regards,
Amit

[Ganesan R · Accepted solution]

Hi,

>>>i wanted to know that why on some setup it writes in file like "lan0" only
and on some setup it writes like "lan0,lan3".<<<

Take the first one..

# Warning: There are no standby network interfaces for lan0,lan3
# Warning: There are no standby network interfaces for lan1,lan2.

Here lan0 and lan3(or lan1 and lan2) might be in same subnet and do not have standby ethernet cards. Means there is no ethernet card connected to that subnet without ip configured.


Another setup: -

# Warning: There are no standby network interfaces for lan0.
# Warning: There are no standby network interfaces for lan1.
# Warning: There are no standby network interfaces for lan2.

Here lan0, lan1, lan2 might be in different networks and do not have standby cards on the network.

[astra]

Hello Ganesan,

I find the following in netconf file: -

The subnets of all the network inetrfaces are different. Please see the following: -

INTERFACE_NAME[0]="lan0"
IP_ADDRESS[0]="11.130.247.1"
SUBNET_MASK[0]="255.255.255.224"
BROADCAST_ADDRESS[0]=""
INTERFACE_STATE[0]=""
DHCP_ENABLE[0]=0
INTERFACE_MODULES[0]=""

INTERFACE_NAME[1]=lan2
IP_ADDRESS[1]=172.27.10.2
SUBNET_MASK[1]=255.255.255.0
BROADCAST_ADDRESS[1]=""
INTERFACE_STATE[1]=up
DHCP_ENABLE[1]=0

INTERFACE_NAME[2]=lan1
IP_ADDRESS[2]=192.168.0.1
SUBNET_MASK[2]=255.255.255.0
BROADCAST_ADDRESS[2]=""
INTERFACE_STATE[2]=up
DHCP_ENABLE[2]=0

INTERFACE_NAME[3]=lan3
IP_ADDRESS[3]=192.168.10.69
SUBNET_MASK[3]=255.255.255.240
BROADCAST_ADDRESS[3]=""
INTERFACE_STATE[3]=up
DHCP_ENABLE[3]=0

If it is true then why it is coming like

# Warning: There are no standby network interfaces for lan0,lan3 <===
# Warning: There are no standby network interfaces for lan1,lan2 <===

and not like

# Warning: There are no standby network interfaces for lan0
# Warning: There are no standby network interfaces for lan1
# Warning: There are no standby network interfaces for lan2
# Warning: There are no standby network interfaces for lan3

Please clarify?

Regards,
Amit

[Ganesan R]

Hi Amit,

Well. From netconf, lan1,lan2 in one network and lan0,lan3 are in different subnets.

How about Service Guard version? Are they same on "one setup" and "Another setup".

It may possible that, we get this kind of if versions are different. I don't say the reports are incorrect. But the way it reports could be different on versions.

What we actually need to look at is, the report which it provide is correct or not.
Like, Do we have standby for lan0,lan3 and lan1,lan2 ?or don't we?

Same way, Don't we have standby for lan0,lan1,lan2 on second setup?

[astra]

Hello Ganesan,

Thanks for your reply.

Could you please elaborate how did you identified that the pair of Lans (lan0,lan3) and (lan1 ,lan2) are in the same subnet.??

Thanks in advance.

Regards,
Amit

[sujit kumar singh]

Hello Amit,

might be not in connectiion to the latest query of yours, but i shall tell you to diagnose the errors as they are coming on.
More than the IP address assigned to the LAN I/fs what shall be of more importance shall be like How are the LAN CArds wired to whicg switsches in whicH networks.
if you can post for both the nodes:
ioscan -fnClan
netstat -in
netstat -rn
lanscan
for both the ndoes along with the scereen O/p when you run the cmquerycl and the corresponding /etc/cmcluster/ then it shall be more clearer.
and also as need to have a clear picture How the LAN CAbles are going to Which Switches and in what networks.

regards
sujit

[Ganesan R]

Hi Amit,

>>Could you please elaborate how did you identified that the pair of Lans (lan0,lan3) and (lan1 ,lan2) are in the same subnet.??<<<

I mean to say lan1,lan2 are in same network since it's subnet mask is 255.255.255.0.
And lan0 and lan3 are in different networks since it is subnetted.

That is why I informed you to check the SG version on both setup. Because different SG version may report the results in different ways. Also, as I mentioned earlier we need to look at the result. The way it format and produce the result may not helpful for us.

[astra]

Hello ganesan,

Please find the cmquerycl result.

Looking for other clusters ... Done
Found 30 devices on node gmlc01
Found 30 devices on node gmlc02
Analysis of 60 devices should take approximately 7 seconds
0%----10%----20%----30%----40%----50%----60%----70%----80%----90%----100%
Found 3 volume groups on node gmlc01
Found 3 volume groups on node gmlc02
Analysis of 6 volume groups should take approximately 1 seconds
0%----10%----20%----30%----40%----50%----60%----70%----80%----90%----100%

Node Names: gmlc01
gmlc02

Bridged networks (local node information only - full probing was not performed):

1 lan0 (gmlc01)
lan3 (gmlc01)

2 lan1 (gmlc01)
lan2 (gmlc01)

5 lan0 (gmlc02)
lan3 (gmlc02)

6 lan1 (gmlc02)
lan2 (gmlc02)

IP subnets:

IPv4:

10.129.246.128 lan0 (gmlc01)
lan0 (gmlc02)

192.168.1.0 lan1 (gmlc01)
lan1 (gmlc02)

172.27.120.0 lan2 (gmlc01)
lan2 (gmlc02)

192.168.0.64 lan3 (gmlc01)
lan3 (gmlc02)

Hope this enable to identify something?

Thanks in advance,

Regards,
Amit

[Steven E. Protter]

Shalom Amit,


10.129.246.128 lan0 (gmlc01)
lan0 (gmlc02)

192.168.1.0 lan1 (gmlc01)
lan1 (gmlc02)

172.27.120.0 lan2 (gmlc01)
lan2 (gmlc02)

192.168.0.64 lan3 (gmlc01)
lan3 (gmlc02)


Lan detection appears to be complete on all nodes.

I'm still not after reading through this long thread which node the cmquerycl was run on.

Lets run it on both nodes and post the results.

I suspect a netmask or network hardware (wire or switch) issue involving some of the NIC cards on the node that is producing these results:

# Warning: There are no standby network interfaces for lan0,lan3
# Warning: There are no standby network interfaces for lan1,lan2.

Insure please that the cluster configuration file is the same on both nodes and before starting, do a telnet or ssh and ping test to insure connectivity between all 4 lans on all for NIC ports, both directions both systems.

SEP

[astra]

Hello All,

I got the logs from customer site, i am attaching the following logs: -

- Cmquerycl logs from both the nodes
- lanscan
- ioscan -fnC lan

Ping/ssh/telnet is working fine from both nodes to each other.

Hope this will lead us to some resolution.

Regards,
Amit