I have an issue I have noticed on the past several software updates. The symptom is that if the hub has been running a while on the prior release and I initiate a software update the download goes very slow (15-20 minutes) and eventual the update fails and stops.
Today I was doing an update and the obvious slow download started so I cancelled it and rebooted the hub then immediately tried the update and the software downloaded in less than 60 seconds. This is twice I have seen and recovered from this slow download/failed update by doing a reboot then immediate update.
So I have a work around for now but something is going on in the system that updates are slowed to a crawl and eventually time out.
I have 200+Mbps internet and Hubitat is ethernet tethered to my cable modem.
A secondary problem is when I reboot and get the update going well it completes fine but the web gui gets stuck at the Extracting update... stage. If I come back to the hub with another device or browser I find the system has updated and rebooted. So also more of a nuisance than a serious problem but I wonder how many people are staring at their screen waiting for the update to complete when in fact it already has...
p.s. doing all of the above via remote access hundreds of miles away...
I'm pretty sure there is a Community version of this.... If he hasn't already, I'm sure @thebearmay will have developed one by the time I finish fumbling my way through typing this response... Or he'll know the alternative I have forgotten....
I use a Node RED option to report this on a Grafana dashboard.....
If you disable autonegotiation, it hides link drops and other physical layer problems. Only disable autonegotiation to end-devices, such as older NICs that do not support autonegotiation. Do not disable autonegotiation between switches unless absolutely required, as physical layer problems can go undetected and result in spanning tree loops.
Has anyone actually gotten that speed test in the network tests to work recently? I tried it the other day to show an example for someone but it never actually shows the results for me. Tried it on two different hubs.
Cannot use auto. When in auto and the router is rebooted the hub loses cloud connnection (another bug) even though it really hasnt. Ie ping traceroute dns lookups etc all work its just the cloud dashboards and other cloud urls that fail.
And yes speedtest is not working anymore.
My guess is some thread (that listens or maintains a.socket for incoming cloud connections) dies on the hub when it gets notified of lost connection and it is never restarted/reset except at boot. Setting to fixed the exchange doesnt happen and the thread is not killed or the connection stays up?
I have never had that issue on any of my multiple Hubitat Elevation hubs when using my UniFi network switches and router/gateway. I can reboot any piece of my home network (either via software or physically by pulling the power cable) and my Hubitat hubs always reconnect to the Hubitat cloud. All are set to AUTO negotiate. Thus, if there is a 'bug', it does not affect all users.
Sounds like a hardware issue on one of your hubs, possibly. It does not sound like a software bug, at least not to me. Are all of your 4 hubs the same generation of hardware? Is the one hub with the issue different from the rest in terms of model number? The original C-3/C-4 hubs were definitely different underlying hardware platforms. The C5 through C8 series is Hubitat's original design, and all share many things in common.
