I have a stack of 4 MS250 switches running the latest firmware. The switches are in standard stack config from switch 1 down to switch 4. Switch 1 has an uplink to WAN provider and switch 2 has the back up incase switch 1 loses its WAN link and this is set to failover using VRRP on the WAN provider kit.
I wanted to move the management IP of my switches over to a new subnet so i began with switch 1 which worked fine. I then started to move down the stack to switches 2, 3 and 4 where i ran into issues with switches getting bad DNS and 2 of then losing connectivity all together.
The switch stack holds some host servers so it will need done out of hours to avoid impact. Is it just a case of logging in and reloading switch 1 from the console or will i need to reload each switch or all at the same time?
Another thing that always confuses me is, meraki support know which switch is the master switch but there is no way to get this info from the web portal. Im assuming that to reboot a remote stack i would need to know which switch is master.
@SMS-PLC : Switch with the lowest MAC address takes master role. Do one thing, remove all switches from the stack expect the master which is working fine and factory reset all the other switches and then add them again in the stack.
Thanks @Inderdeep ill try that after a soft reload tonight.
Last time something like this happened we had to gain local access via the management port at the back and add the management settings manually. If that did not work then pulling power from all stack members and doing a hard reload worked.
@SMS-PLC we always set switch (and AP) management IPs to DHCP now. It avoids all of the issues when renumbering or moving them to a different site. Unless you use some traditional network monitoring tools or add ons that need the device to have a static IP then I haven't found a downside.
DNS is outside because at the time we deployed them we didn't have an available internal DNS server.
There is not bad DNS or DNS misconfigured messages on random switches despite them all using googles name service (8.8.x.x).
I think ill be getting local access to them later and taking a look to get the info straight from the horses mouth.
@cmr @PhilipDAth is there a way to keep the stack as is and master reset the problem members? i was thinking that i could reset every switch except the master whilst keeping them in a stack and then rebuild if needed?
I've never had to touch these things in anger before when it comes to makor issues so please excuse the silly questions.