On Mon, 25 Oct 1999 tkaczma@gryf.net wrote:
I think it should be:
nfs1> ifconfig VIF_Subnet1 10.35.8.1 netmask 255.255.255.128 partner VIF_Subnet1 nfs1> ifconfig VIF_Subnet2 10.35.8.129 netmask 255.255.255.128 partner VIF_Subnet2
and:
nfs2> ifconfig VIF_Subnet1 10.35.8.2 netmask 255.255.255.128 partner VIF_Subnet1 nfs2> ifconfig VIF_Subnet2 10.35.8.130 netmask 255.255.255.128 partner VIF_Subnet2
Right, and I think Sam's followup post explains how this works. If nfs1 fails, nfs2 will see that the partner trunk for the first interface is called "VIF_Subnet1", and will look up the corresponding ifconfig statement in nfs1's /etc/rc to map it to an IP address. To use a less ambiguous example:
nfs1> ifconfig Toronto 10.35.8.1 netmask 255.255.255.0 partner Ottawa nfs1> ifconfig Montreal 10.35.8.129 netmask 255.255.255.0 partner Vancouver nfs2> ifconfig Ottawa 10.35.8.2 netmask 255.255.255.0 partner Toronto nfs2> ifconfig Vancouver 10.35.8.130 netmask 255.255.255.0 partner Montreal
This is because I think that you cannot use IP addreses with trunk interfaces.
That doesn't seem to be the case, since I can ifconfig an IP address to a trunk group. The benefit, as I see it, is that it forces the specification of the IP address to a single location. There is less chance that you'll make an IP typo in the "partner" statement. Also, if you change the IP address on the active filer's /etc/rc, the partner filer automatically knows about it, without requiring additional changes to its /etc/rc.
I have no idea, but I'd skip underlines for legacy purposes. NACs try to resolve the NIC name on takeover i.e. they will try to look up VIF_Subnet2.your.domain.
Oh? I don't think that is mentioned in the manual. VIF names need to have corresponding IP addresses in the filer's /etc/hosts then?