Is this a good rationale/design?

[news2010a]

Imagine you need to provide Cisco switches for manufacturing equipment (i.e., robots). This is a mission critical environment.

Instead of placing (2) stacked Cisco 3750 48 port switches (with RPS 2300) I would put (2) 3750 24 port switches. That way I would force low density thus minimizing the impact on the manufacturing facility in case switch goes down or suffers interruption.

Does that rationale make sense to you?

[Danilo Dy]

Hi,

Yes.

However, take note that there is some downside of having STACK switch...

1. During IOS update, all switches in the STACK need to reboot together. This cause a total downtime, and if in case you encounter problem during reboot, the downtime will be longer.

2. In a L3 STACK configuration, if the MASTER switch fails, all interface MAC address change. There are systems that could not learn this change automatically.

If uptime is important, I suggest not to use STACK and also find a switch with higher MTBF and use GLBP or HSRP. Make sure that RPS is connected to two different power source (PDU).

It is important also to have a maintenance window. Monthly, Quarterly, Half-yearly. To use for IOS update, testing the failover/DR, etc. than to wait for something to breakdown and when that happens your supposed to be redundant setup may not work. Sure you can test it before production, but along the way something wear and tear.

Regards,

Dandy

[Joseph W. Doherty]

"2. In a L3 STACK configuration, if the MASTER switch fails, all interface MAC address change. There are systems that could not learn this change automatically."

One possible way to avoid this is usage of HSRP within the stack. Another, with 12.2(35)SE and later, stack-mac persistent timer 0.