RPi B+ GUI Dies. SSH OK Till Reboot. Then Nothing.

[iceguy]

Hardware: RPi B+ and HiFiBerry DAC+. NAS on WIndows 7 Box.

Software: Rune 0.3 with updates as of June 15.

Problem: Everything worked fine for a month or so. Then the GUI stopped responding.

What I tried: At that point, I could still login with ssh. I used ps to find that mpd was running.

I wanted to restart mpd and/or the web server, but I didn't know how to do that.

I probably should have used shutdown -r now to reboot the system, but I just pulled the power plug waited a few seconds and reconnected it. After a couple of minutes when the RPi appeared to be back up, I tried to reach the GUI through a browser and failed. Then I tried to reach the RPi via ssh. That failed, too.

This has happened a few times before -- though I didn't try getting through with ssh before rebooting.

All I can do at this point is reinstall, update, and reconfigure, about half an hour's work. Ugh.

And I have no idea what caused the problem.

Suggestions?

[Stymieddd]

I have the same problem. If the power source flickers or I unplug the device, I have to reinstall.

[s8ntandrew]

+1. Same issue here.

[hondagx35]

Hi all,

If the power source flickers or I unplug the device, I have to reinstall.

This shouldn't happen all the time, but it is quite common on devices running from sd cards.
This is nothing, that is related to RuneAudio.

The team is working on a solution for that, but it is not an easy task.
The solution for this is to get the filesystem read only.

Frank

[Stymieddd]

Thank you.

[iceguy]

I doubt that power flickers are the cause of the GUI dying.

I own quite a few 1500 Amp-hr UPS's. I put one behind my home theater/audio system and plugged the RPi B+/HifiBerry into it along with my satellite receiver, the WiFi network router, and my Sony TV into that unit. The UPS told me that it could run hours with that load.

I reflashed the memory for RPi B+/HifiBerry and went through the rather tedious setup procedure. I tested runeaudio and found that it all worked fine.

The next night I was watching satellite TV when the line AC went out for a full second and then came back. There was no interruption on the TV, though my Onkyo receiver handling the audio clicked off. Then the line AC then went out and returned twice more in quick succession. Again the TV and video signal went right on.

Using my web browser on my smartphone, I checked the GUI and found that it was working fine! Clearly it had survived the three power outages. After restarting the audio receiver, I could play music from my NAS -- on a similar UPS in another room -- without difficulty. Clearly the UPS kept the RPi setup running through the power outage.

For a couple of weeks I used the RPi B+/HifiBerry DAC through the rune GUI, but one day I opened the GUI from my smartphone and it didn't work anymore. While it could see the RPi B+/HifiBerry DAC, I could not get it to add music to the queue, whether I tried to do it from a playlist or from the library itself. I checked the NAS and it was running fine.

The GUI was locked up.

So the problem is not just power related. I suspect that it occurs when I leave the GUI connected to some browser on a machine that goes to sleep, and then try to re-access it from another browser on another machine.

Also, this time, I ssh-ed in to the RPi as root and issued a "shutdown -r now" to reboot. After a few minutes, I could ping the RPi B+/DAC at the same static IP address I had given it during the last setup. However, the GUI was not up and wouldn't start. Neither could I ssh in. Clearly the ssh daemon was not up and running.

It would be nice to be able to restart the system in this circumstance, rather than reflashing and going through the setup, but I suspect that rune does not include the commands to start the appropriate daemons on boot, including those that bring up ssh and the web server to restart the GUI, typically found down some path like /etc/rc.d/init.d.

[neanderthaler]

ummm...
You won't find any rc.d on rune: Arch (on which rune is based) uses systemd.
So systemctl will be your friend. Try

Code: Select all

systemctl --help

or just

Code: Select all

systemctl

for a list of all services.
As an example, try

Code: Select all

systemctl restart sshd.service

Guess what!

[hondagx35]

Hi all,

Klaus is right "systemctl" is the way to go.

Code: Select all

systemctl restart rune_SY_wrk

This is runes main worker service.

Code: Select all

systemctl restart nginx

The webserver

Frank

[iceguy]

Thanks for the tips on restarting the Unix services. I'll keep those on the tips of my fingers.

I have discovered that what I thought was runeaudio hanging is probably mpd being extremely slow to respond!

If I tell runeaudio to ask mpd to play the queue after it has idled for a long time, it may do nothing for 10 minutes or more... and then start playing!

I don't know why that would happen, but I suppose it might be rebuilding the library info from the NAS.

In any case, except for that one problem, runeaudio and mpd seem to be working fine for me. Keep in mind that I have left the UPS in place.

For those of you experiencing the "Hung RuneAudio" problem, try a little patience!