This past May, we replaced an old Cisco switch between our Dell TOR switches (2) for our VXRail AND our Meraki MS225 4-Switch stack for the data center. We replaced it with an MS 390-24 port with 8 port 10g module as recommended by the tech team at Meraki. We actually purchased two; one for production and the other for a spare. Currently, both are active in layer 2 with one doing all the lifting and the other just sitting there online.
The 4 fiber trunk from the 2 Dell Switches to the new MS390 is set up with VLANS 5,10,25,40,45,100,160,200-201,3939. Some of the unused copper ports are still set to Trunk, VLAN 1-1000. There is two fiber LACP trunk back to the MS225 stack with the same VLANS.
Several weeks ago we scheduled a firmware update for the Meraki Switches in all of our buildings. When we got to the datacenter where the MS390s live, the primary MS390 got stuck in a boot loop and indicator light stayed amber. The spare MS390 seemed okay. SO, we rolled the firmware back and things stabilized at 16.8 (We were on 16.7 before). once the switches were back up all of our VMs were unavailable on our network. After reconfiguring the LACP ports to the Dell it appeared that VMs began coming back online again. Things are trucking along now, but I want to learn more about what may have happened to cause the firmware failure. We should have submitted a ticket with Meraki but we didn't.
After the dust settled on this I did some looking around on the spare MS390. I noticed that it would not allow me to set up ports with the same Allowed VLANS as the production switch. It claimed that the switch was limited to only 1000 VLANS and then it listed out the unique VLAN configurations on the switch indicating that was more than 1000, and it is. So it made me question how production switch was set up violating an apparent limitation. When we received the switches I updated the firmware on both and configured the ports on the production MS390 with the needed VLANS in preparation to install with no problems.
Does anyone use the MS390s in the same or similar scenario?
Any issues with firmware updates?
What could have caused the the firmware update failure?
What would cause the VMs to lose connectivity through the updated switch?
Should these switches be on their own Meraki network so the firmware doesn’t have to be updated as often?
Thanks in advance for any commentary or insights into this.
Shane