ROCK - Roon Server Restarting Itself Often

NUC 7i7 120GB/8GB running latest ROCK build 1.7 b667
Storage is via two 4TB USB HDDs (Seagate and WD Passport) plus folders on a Synology NAS (DS416j). This config has been working fine for the last two years, approx 270,000 tracks. NUC has been running ROCK for over 3 years.

Router is ISP supplied Virginmedia Superhub 2. Network is multiple Netgear Gb connected unmanaged switches (GS108 mainly). NUC and all endpoints but one are wired.

Endpoints:
Wired - Five RPi 3B using IQAudio HATs
Wireless - One RPi 3B using USB connection to Chord Mojo DAC
All Pi are on RoPieeeXL 3.004 (soon to be updated to latest release)

Description Of Issue
Roon Server is restarting itself regularly, the latest being just over a day after a complete restart of the NUC (I restarted Roon OS from the web interface). Previous to that Roon Server would go a few days or a week before Roon Server restarted itself. As the restarts are getting closer together, it is somewhat worrying!

I don’t know when this issue started but it is likely to be in the last two weeks or I would have noticed sooner. During the last two weeks there have been some changes to the system:
Christmas music has been added to storage (both local and from Tidal)
Five Pi 3B endpoints have been updated from DietPi to RoPieeeXL (all wired)
One new Pi 3B endpoint has been added (RoPieeeXL) and it is wireless.
During these updates, the names of some of the endpoints will have changed slightly; just mentioning that incase it is relevant. Also, IP addresses of endpoints have all changed (now on DHCP from being fixed previously).

I think the first Roon Server restart may have happened before I started converting the endpoints from DietPi to RoPieeeXL, but cannot be sure.

I have grabbed copies of the logs from the first two restarts, but can grab a fresh one for the latest restart. If you could provide somewhere to upload them to, I can send them in.

I run a similar rock setup but with 16gb ram and no issues, that said running 8 I didn’t have any issues either.

I was running in a Fanless chassis but if you are running with a fan then check to see the fan blades are clean and no dust in the heatsink fins.

Thanks, I will take a look.

It would be useful if ROCK reported any resource and temperature issues on the web page of course. 8GB was enough over 3 years ago when ROCK was released, but may no longer be.

1 Like

Whilst I did not remove the cover, the NUC is operating at usual temperature (lukewarm to the touch) with all vents clear. The fan is spinning slowly as usual, with lukewarm air coming out of the vents. If it was overheating, I would expect it would sound more like a hairdryer. It lives in the garage anyway, which isn’t a dusty environment, and here in the UK is at about 10 C at the moment, so nice cool air for it.

Let’s see if support will want to turn on diagnostics or get you send a zip up of the logs from rocks roon server folder

My NUC 7i7 (32GB running latest ROCK build 1.7 b667, within AKASA Turing casing, external SSD, ethernet, with Wifi disabled) has never restarted itself.

Nowadays I don’t oversample, but when I had upsampled in the past up to 256 DSD, the processor remained well below the healthy 50 degrees celcius.

Once I had to restart the NUC myself in order to let Roon ROCK recognise Google Home Max (for use as secondary easy listening in the living room); after a restart of the NUC the combination operates flawless, of course within the limitations of Google Home Max.

With the kind of symptom in your description, I start my fault analysis with checking the power supply: are all plugs places correctly put in all sockets. Consider if the main power supply 230 V (or 110 V) may have tiny drop outs, causing an interrupt of the power supply to your NUC.
While placing new gear near your NUC, you may have partially loosened one of these plugs, causing a drop out in voltage once in a while.

Internal software issue in a NUC/server/PC show itself, by freezing screen (part of the software is waiting for another part that unfortunately will never arrive), or by “blue/white screen of death” mostly indicating that the operating system had unjustified overwritten part of the (working) memory and hampering the integrity of the system, or worse: unable to reboot due to lacking pieces of essential software.

Then I would start considering what kind of external cause/signal may let your NUC restart. In general these kind of causes/signals are rare…unless a part of the NUC itself is starting to slowly disintegrate or move outside its specs.

Is it Roon Server (the process) restarting or the NUC?

1 Like

Roon Server only. The Nuc itself does not reboot. Roon OS stays up.

Have you reinstalled at any point in the troubleshooting?

I have to wonder if this is related to your issue.

1 Like

Hi,

No, I am concerned that it may be a resource issue, but have no way of proving it (unless anyone knows what to look for in the logs?). I would rather have a diagnosis than just cross my fingers and do a reinstall. It is an appliance, so should not need it.

Do you have access to system logs on ROCK?

Hi @Anthony_Bates,

I can do this :slight_smile:
I just sent you the location via PM where you can upload the logs!

Do you by any chance recall the time + date of the restart?
It would help us cross-reference the logs.

The server portion has always restarted for me since ROCK was released. It rarely goes past several days of up time. And at times, It would restart 6 or 10 times a day. Never during actual usage though.

So you don’t experience Roon Server running for this long @Rugby?

image

Many thanks, logs uploaded. Restart was approx 24 hours before I collected the logs.

Hi,
I generally restart Server from the web interface every 30 days or so as Roon response time can get very sluggish after this sort of time. But it almost never restarts by itself.
In the past it has been up for over a 100 days without issue, but in the last year or so, I feel there is some sort of resource issue which requires a regular restart every now and then. Perhaps I just need more than 8GB of RAM now, what with the increase in metadata in the DB?

You do have a LOT of tracks :slight_smile:

Upgrading to 16GB isn’t a problem, if that is the solution for this library size, but I would like to know for sure before going ahead.