How to set primary interface in failsafe ip tcpip version 5.6

[Clark Powell]

I don't know how I got into this situation but apparently I either have no primary interface or two primary interfaces for my failsafe ip setup. After failing over from the 1000 Mb interface WE0 to the 100 Mb interface IE0 the failsafe did not move back to the 1000 Mb interface when that line was up.
I've been searching the literature to try to figure out how to change this but have not found the exact command to do it. Anyone know the correct command to create/change the primary interface?


ALPHAD> IFCONFIG -A
IE0: flags=c43
*inet 172.20.66.165 netmask fffffe00 broadcast 172.20.67.255 ipmtu 1500

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

TN0: flags=80

TN1: flags=80

WE0: flags=c43
Standby addresses:
*inet 172.20.66.165 netmask fffffe00 broadcast 172.20.67.255 (Active on A
LPHAD IE0)

thanks
Clark Powell

[Andy Bustamante]

As long as there area active connections, the address will stay on the second interface. You can force it back to the original with:

$ ifconfig ieo fail

$ ifconfig ieo -fail

You may consider assigning IE0 an WEO unique addresses with no failover then assign a third address as the "service" address configured for fail over between the two interfaces.

See http://h71000.www7.hp.com/openvms/journal/v2/articles/tcpip.html for Matt Muggeridge's OpenVMS Technical Journal reviewing failSAFE IP.

[Hoff]

Start reading at:

http://h71000.www7.hp.com/doc/83final/6526/6526pro_006.html

and read and scroll down to the section on how to simulate a failure with the logical name:

TCPIP$FAILSAFE_FAILED_ifname

Check the failsafe log from the activity, too.

And as there have been various issues on various releases, load the UPDATE ECO (and the LAN ECO, if that's not included) and load the TCPIP ECO.

And if you're seeing something weird (and you're current on your ECO kits), ring up HP.

[Matt Muggeridge]

Hi Clark,


>>$ ifconfig ieo fail
>>
>>$ ifconfig ieo -fail

I just want to point out that you should *not* issue this sequence of commands unless you really understand what goes on under the hood. As few appreciate the inner workings of TCP/IP, I regret documenting it.

The problem with these commands is that it they represent just one step in the sequence of failover/recovery. Most importantly, when you issue the "-fail", it does not recover static routes to that interface. You would need to do that manually, amongst other possible steps.

It's far, far, far, better to let failSAFE do the work. As Hoff described:

$ reply/enable ! so you can see the events
$ define/sys tcpip$failsafe_failed_ie0 1
$! wait for failover
$ deass/sys tcpip$failsafe_failed_ie0

If problems persist, and for future reference:

It would also have helped if you listed the output of:

$ ifconfig -av

Also, if you are using your own procedures to create your addresses (as opposed to using TCPIP$CONFIG), then it would help to see those too.

Cheers,
Matt.

[Clark Powell]

OK, maybe the correct term is "home interface." It is identified with an "*" and if I understand correctly the failsafe ip address may fail over to a non-home interface but after the home interface recovers the ip address will move back to the home address automatically, (as long as there are no active connections.) So this brings up two questions:
1.) why does this system have two "home interfaces" for the same ip address? That is, two interfaces with an "*." Isn't that paradoxical?

ALPHAD> ifconfig -av
IE0: flags=c43
*inet 172.20.66.165 netmask fffffe00 broadcast 172.20.67.255 ipmtu 1500 (Sta
ndby on ALPHAD)
HWaddr aa:00:04:00:0b:04

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

TN0: flags=80

TN1: flags=80

WE0: flags=c43
Standby addresses:
*inet 172.20.66.165 netmask fffffe00 broadcast 172.20.67.255 (Active on A
LPHAD IE0)
HWaddr 00:18:fe:83:85:f1

These were created using tcpip$config.com

What I want is to have the interface go back to the 1000 MB interface as soon as it is functional again. My interpretation is that that is what a "home interface" is for, to be a prefered interface in the failsafe ip scheme. But, I have two home interfaces so I need to reduce this to one. But, I've only seen literature for adding a home interface and not for removing that quality from an interface.