Multicast MAC address

[Jim Carter]

Does MC/ServiceGuard 11.13 multicast MAC addresses across the networks?

[melvyn burnard]

Could you elaborate a bit more on the question?
SG (all versions) will re-arp the MAC address of the card which the relocatable ip address moves to, but I am not sure thaht is what you are asking

[Jim Carter]

The real issue deals with CampusCluster. When a package fails from one data center to another, the switches in the new data center need to learn the new IP/MAC relationship. The network manager wants to know how the switches perform this function in a < 5 minute timeframe.

[melvyn burnard · Accepted solution]

This is what I thought you might be asking :-}
When the package ip gets added to a new LAN card, there is an automatic arp rebroadcast to set up the MAC/IP address relationship on network devices. This happens as the IP address is added to the network card, so virtually no delay. Wher ewe have seen problems is wher eolder net devices do NOT accept arp rebroadcasts and only refresh their arp cache after a set period of something like 10 nminutes.
HTH

[Jeff Schussele]

James,

I've seen the exact problem Melvyn describes on older Ciscos with outdated firmware. We've even experienced a greater level of frustration with these units ie they will even literally "lose" the floater IP from their ARP table altogether for periods FAR greater than 10 minutes - sometimes until you reassert the IP.
Only solution was to upgrade the Ciscos.

RGds,
Jeff

[Jim Carter]

Jeff,
Can you tell me if you experienced this issue with the Cisco 6900 series?

[Jeff Schussele]

James,

No, we are a large ATM/SONET based site & we were having poblems with the 12000 series switches.
These are our main switches.
But then again we had trouble from the get-go w/these switches as not all vendors are on the same page when it comes to ATM standards & we had to have a senior Cisco engineer come in & rewrite the firmware at install. So we were not sure from where our problem stemmed. This is probably why the net group was reluctant to remain current - we finally forced them to come up when the problem became untenable
We are a mixed host (SUN & HP) shop & this problem was only in our larger HP (V & K-class) & only affected the floater. This problem did not affect SUN's HA product, though - just HP MC/SG.
We always also had problems with situations when the switches would lock up or go 100% CPU such that no traffic would move. If this occurred & lasted longer that 30-45 minutes then we would HAVE to reboot the HPs & SUNs (ATM only) to clear the problem. Thank god we were a testing lab & not production....

HTH,
Jeff