Stack Manager - 1+1 vs N+1 mode

[xylofi]

Hey everyone, 

Is anyone able to clarify what the benefit of using the following commands are, and what exactly are they doing?

switch 1 role active
switch 1 role standby

1:1 redundancy is used to assign active and standby roles to specific switches in the stack. This overrides the traditional N+1 role selection algorithm, where any switch in the stack can be active or standby.

What exactly is the active switch managing? I've been getting this confused with setting the Switch member priorities.

Any insight would be greatly appreciated.

[balaji.bandi]

if this related to Cat 9300 and (may be same for Cat 3850)  - this more related to power.

Fully redundant power solutions (1:1) are often underutilized because every switch has a backup power supply that is idle under normal conditions. However, partially redundant power solutions (1:N) take time to come online in the event of a power failure leading to an outage.

check the below white paper which explain between both :

https://www.cisco.com/c/en/us/products/collateral/switches/catalyst-9300-series-switches/white-paper-c11-741945.html

[xylofi]

Thank you @balaji.bandi this is helpful.

It does look like it is referring to that, shame the white paper doesn't have command examples. I also can't find them in the command reference manual. I'd appreciate clarification if possible, it looks like you're correct but I'd like to be 100%.

The reason I'm asking is because we've ran into this issue recently, https://www.cisco.com/c/en/us/support/docs/switches/catalyst-9300-series-switches/212912-catalyst-9000-switches-booting-to-switch.html

Appreciate the help!

[balaji.bandi]

Switch priority and switch number you referring for the Stack to elect which one to be master.

you can refer this guide :

https://www.cisco.com/c/en/us/products/collateral/switches/catalyst-9300-series-switches/white-paper-c11-741468.html

The one i was referring stack power.

The reason I'm asking is because we've ran into this issue

To be honest we have 100s of stacks, never come across those issue you mentioned in the document as long as the configuration deployed as expected. i have many test scenarios after stack build to remove power and stack cable, never ran issue of that.

Hopefully you may come across mentioned issue, first option check the configuration. and make sure you have stable latest code running.

[Joseph W. Doherty]

"What exactly is the active switch managing?"

The active switch supports the management and control plane functions of all the stack member switches.

Individual switch member support their own data plane functions.

[xylofi]

Hey Joseph,

Are you sure that this command is actually for that?

This is where I'm getting confused. I thought what you're describing is set by the #switch <number> priority <number> command. Which elects Switch active and standby roles.

E.g. 

9300#switch <number> priority 15

!Set priority 15 to elect switch in ACTIVE role

9300#switch <number> priority 14

!Set priority 14 to elect switch in STANDBY role

9300#switch <number> priority 13

!Set priority 13 to elect switch in next STANDBY role

9300#switch <number> priority 12

!Set priority 12 to elect switch in next STANDBY role

Thanks!

[Joseph W. Doherty · Accepted solution]

"Are you sure that this command is actually for that?"

The specific commands, you asked about, change how the active switch is selected and what switch takes over as active if the current active switch fails.

In your reply to me, those commands are used to set priorities in the N+1 (traditional) approach.

The new command changes redundancy to an explicit active/standby assignment approach; 1:1 redundancy.

I.e.:

switch switch-number role { active | standby} Example: Device#switch 1 role active

Changes stack mode to 1:1 mode and designates the switch as active or standby.

Information About 1:1 Redundancy

1:1 redundancy is used to assign active and standby roles to specific switches in the stack. This overrides the traditional N+1 role selection algorithm, where any switch in the stack can be active or standby. In 1:1 redundancy, the stack manager determines the active and standby role for a specific switch, based on the flash ROMMON variable. The algorithm assigns one switch as active, another switch as standby, designating all remaining switches in the stack as members. When an active switch reboots it becomes standby and the existing standby switch will become active. The existing member switches remain in the same state.

That said, to your question what the "active" switch manages, is what the active switch in the N+1 Redundancy model manages.

As to the benefit of this feature, I guess it just precludes all the other stack members from becoming the stack master.  Personally, I don't see that as a significant benefit, but then I never thought set stack master election priorities very important either.