Failover IP addresses

[matthew robey]

OVms V7.3-2 Tcpip V5.4 Eco 4
Two node ES47 Cluster.

I have configured IP failover for two nics on each node of the cluster.
All works well, however when rebooting the cluster, one node displays the following error message:

%TCPIP-E-INVINTER, error defining interface: WEA1
-TCPIP-I-DRIVERQIO, interface operation from internet driver QIO failed
-SYSTEM-F-DUPLNAM, duplicate name
%TCPIP-E-INVINTER, error defining interface: IEB1
-TCPIP-I-DRIVERQIO, interface operation from internet driver QIO failed

The other cluster node displays no such error.

On checking the config all appears to be ok.

$ifconfig -a

IE1: flags=c43
failSAFE IP Addresses:
inet xxx.xxx.xxx.157 netmask ffffff00 broadcast xxx.xxx.xxx.255 (on FIRE7 WE0)
*inet xxx.xxx.xxx.167 netmask ffffff00 broadcast xxx.xxx.xxx.255 ipmtu 1500

LO0: flags=100c89
inet 127.0.0.1 netmask ff000000 ipmtu 4096

TN0: flags=80

TN1: flags=80

WE0: flags=c43
failSAFE IP Addresses:
inet xxx.xxx.xxx.167 netmask ffffff00 broadcast xxx.xxx.xxx.255 (on FIRE7 IE1)
*inet xxx.xxx.xxx.157 netmask ffffff00 broadcast xxx.xxx.xxx.255 ipmtu 1500


Network appears to be ok, any ideas ?

Thanks,

Matt

[John Gillings]

Matt is the tertiary message for IEB1 also DUPLNAM? Or is it really missing as posted?

What does SHOW DEV/FULL WEA and IEB show? How many other protocols are running on the devices?

[John Gillings]

If that doesn't help, please log a case with your local customer support centre.

[matthew robey]

Hi John,

yes, sorry the "duplicate name" for IEB1 is also displayed. (my cut and paste is not what it should be!)

These nics are only running ip and decnet over ip.

Matt

[Volker Halle · Accepted solution]

Matt,

is it possible, that you have (still) defined pseudo interfaces with TCPIP - in addition to failSAFE IP - on that one node ?

Check with TCPIP SHOW CONF INTERFACES

Maybe those definitions conflict with the failSAFE ip config ?

By just defining failSAFE IP, you don't get pseudo interfaces definitions (i.e. weA1).

Volker.

[Robert_Boyd]

In addition to checking the TCPIP interface config info, verify that VMS and LAN databases are in agreement:

$ show dev ie0
$ show dev ew0
$ mcr lancp
lancp> list dev/char
lancp> show dev/char

Also at the console level, verify the device configuration with commands like
>>> show dev i
>>> show dev ew
>>> show *mode

Make sure that there is consistency between the various addresses/settings for the interfaces.

Robert

[matthew robey]

Volker,

that was it ! Thanks for your assistance.

Robert, I'll check the console level stuff next time the system is rebooted. (Production server, could be 6-12 months !)

Matt