My vyatta test setup includes two identically equipped older athlon xp systems where eth0=onboard nforce, eth[1-3]=r8169 based cards. Everything is working fine on both systems, but this weekend I spent about an hour trying to get VRRP to work for fail-over. It works fine on eth0 (onboard nforce) but I couldn't get it to work on eth1-3. In exploring the issue, it appears that not all drivers support the ability to set the MAC address (which it appears VRRP needs). I found the following post:
http://www.kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.22-git1.log It appears to indicate the r8169 driver didn't get the ability to set its MAC address until sometime in kernel 2.6.22 which obviously does me no good at the moment. This isn't a big deal financially, as the only reason I bought the cards was that Fry's had them on sale for $4.99 each and they had a low profile bracket which fit the cases I was using. However, it might be useful to put a blurb in the VRRP section of the documentation stating that the card's driver must support setting the MAC address for VRRP to work (and maybe even list which drivers support and don't support it although I can see how this list might be difficult to compile). FWIW, I also notice in: https://bugzilla.vyatta.com/show_bug.cgi?id=2370 that the latest greatest build has support for a disable-vmac option - but when I tried it in VC3 I just got syntax errors. I'm assuming this would fix the problem also as the card then wouldn't have to set its MAC address but just use it as is? How hard is it to upgrade to a nightly build (we're still a few months away from production so I wouldn't be too concerned with stability)? Any suggestions other than use a different card? Thanks, Jeff _______________________________________________ Vyatta-users mailing list Vyatta-users@mailman.vyatta.com http://mailman.vyatta.com/mailman/listinfo/vyatta-users
