MS250 - Stacked Switches - Issue after reboot

[Dougbert107]

Greetings,

Wondered if anyone else has experienced this with the MS250-48 switches when stacked. We have two MS250 (CSW-01) switches stacked together via the stacking port. Last night we rebooted both (first CSW02 and then CSW01) after hours and when we came in this morning ever other packet was dropping to our firewall and to 8.8.8.8. 

The stack is configured with:
L3 interfaces for our VLANs

DHCP services for the VLANs

Single static route (0.0.0.0/0.0.0.0) to the firewall. 

No routing protocols running.

We power cycled CSW02 and once it came back online everything started working normally. I can't find much in terms of documentation, not sure if CSW02 was trying to act as the master between both switches because it it's uptime was longer than CSW01. I have a ticket in with support but I was curious if this is normal behavior or if we should make sure the CSW01 switch gets rebooted first before CSW02 to avoid this issue.

[cmr]

@Dougbert107 with Meraki switches you should always reboot the stack as one, unlike Cisco IOS there isn't a predefined stack master etc. 

[DarrenOC]

Hi Charles, not sure that’s what happens for MS250 stacks. It appears to be the default behaviour for MS390’s (snippet is taken from section regarding stacking MS390’s)

“Rebooting a stack member (from dashboard/physical power cycle) will reboot all the members in a stack. The same applies when performing factory reset on a stack member. “

https://documentation.meraki.com/MS/Stacking/Switch_Stacks

ive rebooted plenty of 250’s and not seen the other stack members reboot nor had to bounce the others.

@Dougbert107 - is your Stack cabled correctly? Have you tried reseating the stack cables?  Are you on the latest firmware?

[cmr]

@DarrenOC I was going off our experience with the 355s where you have to cable them up and then power them on for them to stack properly (unlike Cisco IOS where you can add online), so I thought that the same may well be true for rebooting (as Cisco IOS switches behave nicely in this regard).

Could be wrong, but the warnings about not choosing stack members when delaying an upgrade make me think it isn't advised.  

But as you've not had issues with 250 stack member reboots, my summations were obviously not the case 🤦‍♂️

Hey ho 😌

[DarrenOC]

I could bounce some 250 stacks now and see how they behave 😁

Might get a slap though

[cmr]

Go on, you know you want to (and everyone will be asleep), maybe...

[Dougbert107]

Thanks for the reply. I have confirmed the stacking cables are seated correctly. Meraki support couldn't see any issues (and subsequently told me they couldn't pull any logs after the second switch was rebooted).