Switch LAN Failed

[Marcelo Junior_1]

Hi Guys!!

I have a cluster with a package of a Database Oracle. Both 2 nodes of the cluster have a Primary lan ( lan1 ) and a Standby ( Lan2 ).In the package the feature LOCAL_LAN_FAILOVER_ALLOWED have the value yes but if the lan1 failed the lan2 dowsn1t work. In the syslog.log i have this error:

Jan 20 04:38:37 HXMCDB02 cmcld: Failed to switch 172.23.12.213 from lan1(0,0) to lan2(3,ac170c00): Invalid argument
Jan 20 04:38:37 HXMCDB02 cmcld: Failed to switch 172.23.12.215 from lan1:1(0,0) to lan2(3,ac170c00): Invalid argument

I checked the IPs and Subnet from packages an netconf file and this is ok, if I configure manually the IP of package in the lan2 it`s work fine too.

Anybody can help me?

The Service Guard Version is A.11.16.00

Thanks.

[Steven E. Protter]

Shalom,

Check thos configuration files. Something inside there is not consistent with the actual LAN environment.

SEP

[Mark Nieuwboer]

Hi can give your clusterconfig file.
I think there is the problem.

grtz. Mark

[john korterman]

Hi,

if you run hpux 11.00, check also Document ID: 4000068465 in the Tech Base.

regards,
John K.

[melvyn burnard]

Make sure your standby lan is NOT configured, or marked UP, when SG is not running. Also check your config and contriol files to make sure there is not an issue there.
Finally do a cmscancl and look at the output file to determine if SG can see the network connectivity.
Has this worked previously>? If so has anything changed?

[Marcelo Junior_1]

Steven, Mark, Jonh and Melvyn

the system is a HP-UX 11.11

i checked the configurations files, but i don`t find anything wrong, in attachment i put the netconf of both nodes, the scancl.out file and the directory /etc/cmcluster to everybody know the cluster configuration

Tks

Marcelo

PS: i don`t know if this issue worked previously, the first time i need this issue to work is 01/18/2006

[melvyn burnard · Accepted solution]

Your lan2 does not show up in the netstat output on DB02:
------ Output of netstat -in (HXMCDB02) ------

Name Mtu Network Address Ipkts Ierrs Opkts Oerrs Coll
lan3 1500 172.29.7.144 172.29.7.147 1640469368 0 3021862192 0 0
lan1:1 1500 172.23.12.0 172.23.12.215 321975 0 3370 0 0
lan1 1500 172.23.12.0 172.23.12.213 100634884 0 157810220 0 0
lo0 4136 127.0.0.0 127.0.0.1 5916919 0 5916920 0 0

It should be there if it is useable

This means there is a network config issue that needs to be resolved.
Your netconf file looks ok, I would suggest using cmhaltnode amnd cmrunnode to see if that fixes it, or else a reboot of the node

[Lolupee]

It seems you have an issue with the lan2 on db02 as indicated by Melvyn.

you data shows there is connectivity btw LAN1 and LAN2 but netstat cannot see the LAN2. If this is not an error, ten check that out.

the cluster does not recognise LAN2 on db02. can you check you syslog to confirm the error relating to the LAN card during starup of the cluster.

[Marcelo Junior_1]

Melvyn,

i execute the comamnd ifconfig lan2 plumb and now the issue LOCAL_LAN_FAILOVER_ALLOWED work fine.

Thanks for the Help!

Marcelo

[melvyn burnard]

Fixed? Great

[Marcelo Junior_1]

The issue LOCAL_LAN_FAILOVER_ALLOWED is failed why the Standby NIC is not configured. After I execute a ifconfig plumb command in this NIC the feature works properly.

[Lolupee]

Great,

do not forget to look at what might caused the issue for rectification. this is because the issue would come up again when you restart the cluster.

[melvyn burnard]

be nice if he assigned points, too