C8 hub and 2 C7 hubs, all running current release.
I have noticed some rules have become unreliable. Checking the Events tab for the affected devices I see lots of entries of the type:
|hubMeshDisabled|false||Enabled through hub mesh|DEVICE||||2023-03-18 11:27:01.411 PM CDT|
|hubMeshDisabled|true||Disabled through hub mesh|DEVICE||||2023-03-18 11:22:44.088 PM CDT|
For troubleshooting purposes, I'd like to see all events of this type, but I'm not seeing them in the main log output. I have to go to each hub mesh device and open the Events tab.
Is there any central repository of hubMeshDisabled events?
Or, more simply, has anyone else had a similar issue and know a fix? I see at least one similar topic but it's from 2021 and reported as being fixed in 2.2.7.
I have been getting this for a long time, almost since hub mesh's inception. I reported it a long time ago, but it never changes. Doesn't seem to effect anything so I just ignore it.
Don't know if I've been seeing it for a long time, but it has recently started to impact me. If it has happened in the past, the frequency was low enough I didn't notice.
Last night:
Motion sensor (hub mesh) detected motion.
Lights (some hub mesh) turned on.
Mesh disabled, so hub missed the motion stopped notification.
Mesh enabled.
It's happening often enough that other automations are affected as well.
It was a process of getting past a bunch of incorrect assumptions I was making first.
I only noticed it when I upgraded hub1 from C7 to C8, so my first bad assumption was the issue was hub1. It was a bit hard to figure out which hub was having the issue since the only place I could find any output was in hub-mesh shared devices (on the linked side) Events tab. But there were enough non-mesh events that it was hard to see what was happening. So I finally thought to create a virtual switch on all 3 hubs and share them to the other 2. They're not used for anything, so the only entries in the Events tab were mesh disable true/false events. At that point it became clear hub2 was the issue. hub1 <-> hub3 never had mesh events. hub2 <-> hub 1 and hub2 <-> hub 3 had events which were correlated in time.
So now, I assumed it was a hub or network hardware issue. Replace cables, swapped Orbi satellites (hub2 and hub3 both on different satellites) but problem didn't move with network hardware. Upgraded hub2 from C7 to C8 (good excuse to do this as I wanted to anyway - I'm all Z-Wave and had no issues with first upgrade).
Real breakthrough came when I started checking the logs (yes, I know I should have done this sooner) for my Orange Pi server (running Armbian for InfluxDB / Grafan of hub events) connected to the same Orbi as hub2. Network issues in syslog time-coincident with hub mesh issues.
Our house isn't large, but the walls are all stucco with a wire mesh lath. Orbi satellite in question near ground , in a corner, next to a metal filing cabinet full of cabling and a desk with industrial metal legs. I've been using the Orbi system since mid 2020 without issue (that I could detect), so the satellite kept getting relocated over time. Looking back, I think I've been blaming issues from this setup on other things for at least a few months.
So I raised the Orbi higher on the shelf it's been on. Higher that the filing cabinet and desk legs. Speedtest results increased by 40% and the number of hub mesh events dropped significantly. So I might be making another bad assumption, but I'm pretty confident I'll be able to find a satisfactory location that resolves the issue entirely. RB50 Orbi may get the ax for a newer model. Might not fix the issue, but I'm itching to upgrade anyway.
Any improvement was a coincidence and temporary, or wishful thinking on my part.
I was on-line with my phone when another hub mesh issue was on-going. I was surprised to find my phone was connected to the same Orbi satellite as the hub having issues but was experiencing no issues. In fact, it seemed that all WiFi clients of that satellite were operating without issue and 2 of the 3 ethernet connected systems (Hubitat hub and Orange Pi server) were off-line (no easy test for my ethernet connected laptop). This wasn't consistent with my belief that the issue was caused by the backhaul between this satellite and my router.
I googled something like "orbi wifi works ethernet doesn't" and found several reports of certain models of TVs knocking out all ethernet attached devices when users plugged them into an Orbi. Not exactly the same thing, but it seemed like a good lead. Maybe one of my hardwired devices was the issue.
Due to previous troubleshooting, I felt the hub and orange pi were low probability, so I unplugged my laptop and switched it to WiFi. Laptop connects to a thunderbolt hub to a 2.5gbps USB to ethernet dongle to my Orbi (ports are all 1gbps). So lots of places for weirdness or hardware issues.
End result is no hub disconnects for over 60 hours. Previously, I was lucky to go 6 hours and more often issues were every 2-4 hours. So I claim with some confidence that I've tracked down the cause.
I haven't made any attempt to figure out what the exact cause was. In general, the laptop network seemed fine. My intuition is that there's either an intermittent failure with the USB/ethernet dongle or some auto negotiation incompatibility, but we've seen how spot-on my intuition has been to date.
And I'm fully supportive of you upgrading your network, if anyone asks. 
