Self-Reboot Miniserver

[Jedi Tek'Unum]

It appears that some problems, especially network related, require a power cycle to clear up.

Apparently a "Save in Miniserver" does not cause a full reboot.

Has anyone implemented a hardware mechanism to cycle power that can be done remotely?

I'm thinking a timed relay with the Miniserver powered through the NC contacts and the relay driven by a "digital" output. Turning the output on would break the power to the Miniserver and after a few seconds it would be reapplied by the timer in the relay. I believe this would be a "Delayed Off" type of timed relay.

Of course there are many ways to remotely control an AC outlet but those don't work when a UPS is being used.

[Rob_in]

Are you saying a "Reboot Miniserver" via the app doesn't solve your problem? That's very disturbing as power cycle isn't a clean restart and if 'reboot' isn't actually fully restarting all parts of the hardware, including the network interface would say that is a major bug for Loxone to fix ASAP.

Is this to do with your UDP problem (https://groups.google.com/g/loxone-english/c/3eh1vZty6es) that you traced to bad network config?

Robin

[Jedi Tek'Unum]

I wasn't aware of "Reboot Miniserver" in the app. The logic of including it there and not in Config eludes me. I discovered /dev/sys/reboot as well.

The days of clean vs unclean restarts after a power cycle should be over. It's been a long time since I've seen anything corrupted in any device or computer as a result of power loss. This should be especially true of any embedded device like this.

100% confidence in a software initiated reboot is not something I'm ready to apply to Loxone given the history. Plus it is no guarantee that Extensions will be reset. Since this device is now on my list of unreliable devices I think a power cycle is appropriate. So much so that I will likely set up a periodic (monthly?) power cycle as I do for other unreliable devices (cable modem for example). To accomplish that it needs to be via unattended methods.

Yes, this line of concern was triggered by the UDP problem and someone else highlighting that those kinds of problems sometimes require a full power cycle.

While I did have an anomaly in my network configuration it did not affect my UDP problem. After fixing it as far as I could (still incorrect in one of the UI elements) I did a power cycle to be sure. The UDP problem remains and I'm at a dead end.

[Tico]

I’ve had a similar experience of UDP problems that were caused by an ageing SD card. A replacement fixed the issue.

[Rob_in]

On Tuesday, 9 May 2023 at 16:59:25 UTC+2 Jedi Tek'Unum wrote:

The days of clean vs unclean restarts after a power cycle should be over. It's been a long time since I've seen anything corrupted in any device or computer as a result of power loss. This should be especially true of any embedded device like this.

Well, IMHO anyone who thinks they can just pull the plug at any time on a device that has non-volatile storage inside is asking for trouble. While the chance of corruption is much less diminished with modern filesystems it's still there. IMHO you should always perform a clean shut-down or reboot if possible.

 

100% confidence in a software initiated reboot is not something I'm ready to apply to Loxone given the history.

This is super sad given the cost of Loxone gear :(

 

While I did have an anomaly in my network configuration it did not affect my UDP problem. After fixing it as far as I could (still incorrect in one of the UI elements) I did a power cycle to be sure. The UDP problem remains and I'm at a dead end.

Wait - so you are saying your UDP problem persists over not only a software reboot but also power cycle of the entire system? Did you also power cycle your router(s) and the device with which you are trying to communicate?

Robin

 

[Jedi Tek'Unum]

That is bizarre.

Just to rule it out I replaced my ~3 yr old card. It hadn't been heavily written to and was only 1-2% full. Test ran fine. Both old and new are San Disk Industrial 8GB.

Didn't fix my problem.

[Jedi Tek'Unum]

On Wednesday, May 10, 2023 at 12:27:45 AM UTC-5 Rob_in wrote:

On Tuesday, 9 May 2023 at 16:59:25 UTC+2 Jedi Tek'Unum wrote:

The days of clean vs unclean restarts after a power cycle should be over. It's been a long time since I've seen anything corrupted in any device or computer as a result of power loss. This should be especially true of any embedded device like this.

Well, IMHO anyone who thinks they can just pull the plug at any time on a device that has non-volatile storage inside is asking for trouble. While the chance of corruption is much less diminished with modern filesystems it's still there. IMHO you should always perform a clean shut-down or reboot if possible.

In an ideal world that would preferable. I have no idea if its even possible to do a shutdown with Loxone.?

I was a software engineer/architect at a few computer manufacturers and very large software companies for ~40 years developing systems-level software. There is no excuse for corrupt file systems due to power failure in the 21st century. If it happens then the vendor should be out of business.

100% confidence in a software initiated reboot is not something I'm ready to apply to Loxone given the history.

This is super sad given the cost of Loxone gear :(

They are not unique. I agree that quality is below what one would expect given the cost. The world is completely submerged in the era of "good enough" - that isn't.

 

While I did have an anomaly in my network configuration it did not affect my UDP problem. After fixing it as far as I could (still incorrect in one of the UI elements) I did a power cycle to be sure. The UDP problem remains and I'm at a dead end.

Wait - so you are saying your UDP problem persists over not only a software reboot but also power cycle of the entire system? Did you also power cycle your router(s) and the device with which you are trying to communicate?

Yep. My Miniserver, the device I'm having trouble communicating with (a NanoPI NEO2), a ControlByWeb I/O system, and a dumb switch are all powered by my DC UPS. When I power cycle they all do. They are all on the same subnet/VLAN and all are hard coded with IP addresses. I've tried changing UDP port. I've tried using a bogus IP address and then switching back.

Using "nc" (netcat) I can send UDP packets manually with the same info Loxone should be sending and my receiving software works just as it always has.

I've switched to a new industrial SD card. I upgraded to 14.1.x. 

I tried creating a new UDP Virtual Output for testing and had it triggered by a Pulse per Second. Same thing - no packets sent.

The time I've spent struggling with Loxone over the last 5 years would have been better spent developing my own automation system.

[Rob_in]

On Wednesday, 10 May 2023 at 23:05:59 UTC+2 Jedi Tek'Unum wrote:

I was a software engineer/architect at a few computer manufacturers and very large software companies for ~40 years developing systems-level software. There is no excuse for corrupt file systems due to power failure in the 21st century. If it happens then the vendor should be out of business.

Come on now... I've been an IT professional for decades too. We all know how things are supposed to work, and then there is the way things actually work...

 

The world is completely submerged in the era of "good enough" - that isn't.

... especially in that world!

 

 The time I've spent struggling with Loxone over the last 5 years would have been better spent developing my own automation system.

... and especially from a company with such a poor QA record.

Robin