downtime involved when stack member is powered down and powered back

[Rohit Patil]

When I am running 2 switches sw1 and sw2 in stack.

I am continuously pinging from server 1 to server 2 and server 3.

and If I power down the sw1 and power it back again, till the switch reloads and back in service and is member of stack, does it interrupt the services.

Any supporting doc for such scenario.

If there is any method to optimise the downtime.

[Leo Laohoo]

1.  What switch model? 

2.  What firmware are they running on? 

3.  Are the downstream ports configured as Etherchannel?

[Rohit Patil]

It is 9300x-48tx, ios xe 17.12.3. 

Yes LACP port channel/ether channel is running for each server across the physical hardware.

[Leo Laohoo]

1.  Ping the servers. 

2.  Disable one of the ports. 

3.  Does the ping for more than, say, 4, seconds? 

4.  If it does not.  Enable the port and disable the 2nd port. 

5.  Does the ping drop for, say, 4 seconds?

[Leo Laohoo]

---

@Rohit Patil wrote:
It is 9300x-48tx, ios xe 17.12.3. 

---

Post the complete output to the command "sh switch detail".  

[Rohit Patil]

The switch is stacked completely, we see both switches serial numbers.

Also, if all port channels are in up State, and alternate port down does not induce downtime.

[Leo Laohoo]

---

@Rohit Patil wrote:
The switch is stacked completely, we see both switches serial numbers.

---

Post the complete output to the command "sh switch detail".   We want to see the output ourselves.

[Joseph W. Doherty]

"and If I power down the sw1 and power it back again, till the switch reloads and back in service and is member of stack, does it interrupt the services."

It may.  Further, degree of impact can be very variable.

"If there is any method to optimise the downtime."

Sometimes, yes.

"Any supporting doc for such scenario."

Possibly, although you may not find all the information you want in just a single document, nor its significance well emphasized.

[Rohit Patil]

Let me arrange show stack info. I'll email that to you.

[Rohit Patil]

Let me know what all logs u need to determine, what optimisation is needed?

[Joseph W. Doherty]

This isn't a theorical question?

If not, would first like a better explanation of the environment and what you've been seeing.

A couple of important questions, is the problem only when the active/master drops?  Also, are the servers on the same VLAN/subnet or not?

[Rohit Patil]

This isn't a theoretical question, this setup is deployed, and while testing this behaviour was observed.

The setup is on same data VLAN.

I am not sure, but yes in case of master will it take such huge drop? About 46 sec.

[Joseph W. Doherty]

Have you configured SSO mode?

Might be checked by:

show redundancy states

[Rohit Patil]

Let me check and enable.

[Rohit Patil]

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst9300/software/release/17-12/configuration_guide/stck_mgr_ha/b_1712_stck_mgr_ha_9300_cg.html

Let's explore few more options, high speed stacking and sso. 

How can other options be used