Join us for a month-long contest with heaps of swag to win!Learn More ›
I've got 3 MS120-8 switches on my network. Seemingly randomly, one or another will go red in the dashboard and stop responding to attempts to reach local page and the machines behind the switch become unavailable on the network even though they are still running.
They are all running 10.44 firmware.
Has anyone else seen this?
I went through a fairly lengthy process with support trying to troubleshoot which resulted in an RMA. The new switch hasn't gone dark yet, but it hasn't been very long.
One of the reasons for the RMA was that support couldn't really diagnose the problem because the only way to get to the switch's local logs was to power cycle, which get's the switch back running again, while at the same time blowing away the logs (serious flaw, perhaps?).
I really don't want to RMA the other two switches as well, but I just had another one go dark and required a power cycle to get it back. It seems so unlikely that I got three bad switches, but this is pretty frustrating.
Any advice or similar experience would be greatly appreciated.
I've never had that happen with 10.x firmware.
Have you got any spanning tree features enabled that are not default like root guard?
Which MS120-8 do you have ordered? MS120-8P or MS120-8LP
Which devices connected with the Switches?
What ist the DHCP and DNS Server?
The invoice states MS120-8-HW. There is no PoE on the switches.
Behind the one that was just RMA'ed is just an Apple TV and a TV. Those go off a lot, so there are often times when the only device connected is another dumb switch through which the MS120-8 connects to the MX65.
Behind another is a linux server, always up, a couple of laptops (that rarely move) that go up and down fairly regularly, and the same dumb switch as the other two.
Behind the third is a Windows 10 box, usually up, but not always, and the same dumb switch as the other two.
I'm running a Raspberry PI 2 with Raspbian (stretch), isc-dhcp-server, and bind9. Bind9 is configured as local server for the subnet and caching for the internet. Dhcpd updates bind9 for local IPs it's assigns dynamically.
Currently I am not using fixed IPs for these devices. Could that be an issue? Certainly something to try. Maybe it hiccups on lease renewal?
Have you tried connecting the MS directly to the MX to rule out that the problem is your "dumb" switch suffering performance drops?
The logs get blown away on the restart, sadly, so there isn't anything to see. The logs before the issue occur, taken at random times, look just fine.
Do all the routing devices on the network need to support RSTP or STP? Perhaps that's the issue.
The switch in the middle is a Netgear GS108E, which is really not entirely dumb, it's got some configurability, although I leave it stock, don't create any additional vlans, etc. From what I'm reading it may not support STP, the docs are unclear.
For Meraki Switches, RSTP is enabled by default.
For the small Netgear switches STP / RSTP is disabled by default.
I would also enable RSTP on the Netgear switch.