Quantcast
Channel: IEOC - INE's Online Community
Viewing all articles
Browse latest Browse all 10744

GLBP & load-balancing

$
0
0

I have just one more question on GLBP here.  I setup GLBP between r1 and r2 on a network that looks like this:

 

R1 --trunk-- Sw1 --trunk--Sw2 --trunk-- R2

 

I brung the vlan out on Sw1 and Sw2 in SVI's so I could emulate 2 hosts for testing.  So effectively, R1 was 10.0.0.1, R2 was 10.0.0.2, Sw1 was 10.0.0.11, and Sw2 was 10.0.0.12.  The GLBP VIP was 10.0.0.10

I sent a bunch of pings from Sw1 and Sw2 to the VIP and kept checking & clearing the arp cache's on the switches to check I was getting different virtual mac addresses allocated from the AVG for my AVF, and I was.  At the same time, I was doing a #sh ip glbp detail on the AVG to check the "Client Selection Count" was being sequentially incremented between the two AVFs, and it was (i.e. round-robin was working).  The query I have, is that I noticed the Client Selection Count was also being incremented every time I cleared the arp-cache on the switch, and I can't understand why clearing arp-cache would have any affect on GLBP at all? All I'm doing is flushing my local arp-cache on the switch.  Can anyone explain this behaviour?

PS:I tested this multiple times to make sure it wasn't some random co-incidence. And yes, each time I clear the arp-cache, GLBP seems to allocate an AVF for this so called "traffic" I guess.


Viewing all articles
Browse latest Browse all 10744

Trending Articles