Lost Connection....AGAIN

[bills.hancock]

I have 3 C8 hubs

• I have downgraded two of the hubs to the current firmware (....152)
• One of the hubs (for the garage) has been a rock and never dropped from the network
• The other two flake out all the time
• The only way to reconnect is to power cycle

All three hubs are connected via ethernet (all cables have been replaced - used ones that have provided secure connections with other devices)

All have had DHCP reservations from the router
I have tried combinations of Static IP and DHCP from the Hubitat
I have limited connection speed to 100 mbs and allowed for auto
I have tried the auto-reconnect from the network settings (one reason for the down grade to firmware .152)
I have also moved the two hubs around to different locations in the house (maybe wiring in house...although I have interrogated those wires and all connections appear to be working)

I have now used three different routers and 5 different switches

• Currently using an ASUS ET12 router
• Cisco switch

Nothing has worked

Frustration is starting to set in!

Any suggestions would be very welcome

[csteele]

@bills.hancock

• The other two flake out all the time
• The only way to reconnect is to power cycle

That's not a lot of detail to go on...

Is the hub working during a flake out? AS in, are the automations firing at sunrise/sunset, motion is turning on lights, etc and the "flaky" has to do with browser connection to the user interface and or dashboards?

[bills.hancock]

When this started to occur the answer to that question was yes (that was about two weeks ago). The answer to that question is now no.

[rlithgow1]

@bills.hancock

I have tried combinations of Static IP and DHCP from the Hubitat
I have limited connection speed to 100 mbs and allowed for auto

Best settings are DHCP only, and set speed to auto.

Are jumbo frames enabled either in the cisco switch (what model do you have?) or in the asus et12?

Have you done a soft reset?

[csteele]

@bills.hancock

The answer to that question is now no.

That sounds like a corrupt DB and I suggest a Restore of a recent backup. You can make a backup now, and restore that, because the backup/restore process cleans up a corrupt DB.

I suspect the repeated power pulls to get the hub back is the cause of that.

[bills.hancock]

I restored the DB yesterday. Still having the problem.

[garyjmilne]

Setup an outbound ping test from each of the Hubs. Make sure you have enough space in you log. If you have a managed switch I’d pick the up address of the switch to start with.

After it craps out next time take a look at the logs on the two devices and you might be able to draw some useful conclusions.

If you have a desktop you could do some pings test from there redirected to a file in tandem.

[bobbyD]

@bills.hancock

One of the hubs (for the garage) has been a rock and never dropped from the network

This is a good starting point. Can you identify what is different compared to the other 2? Are they all connected to the same type of switch/no switch/access points? Same subnet/different? Same internal network settings for the hubs (E.G fixed 100mbps)?

[jtp10181]

Have you tried to reach the diagnostic tool on port 8081 when they stop working? This will give a clue if the main hub OS has locked up vs a networking issue / disconnect. When the main hub OS locks up usually the 8081 tool is still accessible.

When the entire network interface goes offline, the most common reason in the past has been due to jumbo frames being enabled somewhere on your LAN and a device is sending out multicast jumbo frames which can apparently crash the hub interface.