Hellz ya... I have an extra space in my rack, maybe I'll add another vyatta-duo and do like quad-redundant routing load-balanced bliss!!!
:) Todd Worden Web-Wired, LLC 434.906.0420 [EMAIL PROTECTED] www.web-wired.com -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Stig Thormodsrud Sent: Thursday, January 10, 2008 5:13 PM To: 'Dave Pifke'; 'vyatta-users' Subject: Re: [Vyatta-users] GLBP > > Can canyone comment more on load balaning vrrp? Active/active style > > configuration? Perhaps even noting bgp? I was not aware with vrrp one > > could have two routers handling packets :/ > > This may have changed, but I believe Vyatta only supports one VRRP address > per interface. Consider what I'm describing here a feature request, > although perhaps someone else can comment on how to make this work with > the current functionality. :) Hi Dave, I have already added the support of both multiple vrrp groups per interface and multiple vips per vrrp group in the current development branch. So assuming the testing of these features goes well, then you should see it in the glendale release. I'm hoping to also add support for vrrp sync groups if time permits. stig > If Vyatta supported multiple VRRP addresses (and the equipment behind it > supports ECMP), you could do active/active by configuring two default > gateway addresses and using the VRRP priority/preempt parameters to give > one address an affinity for one router and one for the other. > > For instance: > > Router A, x.x.x.3, VRRP addresses x.x.x.1 priority 100 and x.x.x.2 > priority 50 > > Router B, x.x.x.4, VRRP addresses x.x.x.1 priority 50 and x.x.x.2 priority > 100 > > Device C, x.x.x.5, default gateway configured as x.x.x.1 and x.x.x.2 with > equal metrics > > In normal operation, half the packets will be processed by either router > (depending on how device C implements equal cost multipath). If one > router fails, both the .1 and .2 addresses end up on the surviving box. > N.B. this breaks stateful packet inspection. > > I believe the original reason for the one-addres-per-interface restriction > was due to the virtual MAC address. Now that we have the disable-vmac > option, perhaps this limitation could be removed? > > > - -- > Dave Pifke, [EMAIL PROTECTED] > > -----BEGIN PGP SIGNATURE----- > Version: 2.6.3ia > Charset: noconv > > iQCVAwUBR4aPSTuW2fOIQC3pAQFKmgP/U6kbweEz+HR0Tbrq5aeoXOZu2JXpav4y > fVjBzG8wR7mL/2b1whiVjUq/hj55uiMcXPWQ4+dxWvbRoJgZZx1o1kpjfASW3z+J > aCJ4fbcv0O2fmWqxVGuEc8gPohW3BrBuWOipj1y7vFofmfV7dkEtyOdLLFbaLE9I > Jt7AFqzoFCM= > =ASQ2 > -----END PGP SIGNATURE----- > _______________________________________________ > Vyatta-users mailing list > Vyatta-users@mailman.vyatta.com > http://mailman.vyatta.com/mailman/listinfo/vyatta-users _______________________________________________ Vyatta-users mailing list Vyatta-users@mailman.vyatta.com http://mailman.vyatta.com/mailman/listinfo/vyatta-users __________ NOD32 2781 (20080110) Information __________ This message was checked by NOD32 antivirus system. http://www.eset.com _______________________________________________ Vyatta-users mailing list Vyatta-users@mailman.vyatta.com http://mailman.vyatta.com/mailman/listinfo/vyatta-users
