Build 555 update issue(s)

ROCK running on an i7 Cirrus7 (NUC8i7BEH):

  • Roon did not (auto)restart after applying the update
  • ROCK did non reboot (using web UI, to fix the above and for good measure). had to manually force-quit (NUC was actually stuck on shutting down) then turn it back on

btw: way too many instances of “update now” buttons in the UI (About Roon)
btw2: a ROCK reboot schedule would definitely be a godsend :wink:

another issue: not adding (… yet? used to be immediate!) new Qobuz Favourite Albums to library:

Albums (Library) view:

Qobuz Favourite Albums:

added almost an hour ago and refreshed “My Qobuz” many times. Also tried quitting/restarting Roon

EDIT: had to reboot ROCK to have new Qobuz albums finally show up in Library :roll_eyes:

definitely (… today?) new Qobuz Favorite Albums are NOT being added to my Library :no_mouth:
unless… I reboot ROCK :roll_eyes:

BTW: I’m on a 1GB fiber connection and everything’s working great!

A side question - how has the Cirrus been working? Any problems other than this one?

flawless! :slight_smile:
… 'till these last two Roon builds :roll_eyes:

I wonder how many of your problems related to ROCK come from using fanless chassis.

ROCK relies on the BIOS settings to do thermal management, which in turn use a very effective fan.

If your heat sink is not 100% perfect for dissipating that TDP, you will probably freeze up or have random corruption.

When developing Nucleus, we ran into this a lot, and a thermal management solution in software was required.

and btw, measures and maker’s statements apart… case is not even lukewarm

I would be extremely uncomfortable if I touched the dissipating elements of a passively cooled i7 and found it to be cool. To my mind that would be much more likely an indication of problems.


oh, cmon’: it doesn’t run at even just 10% of its processing power

CPU load on Roon OS sits low until it doesn’t. For example, when you get a metadata update, it’s a pretty significant load on the CPU. The library size and CPU performance will impact how long that load lasts, but it is not minor.

Cirrus7’s testing is good, but the ramp-up seems wrong for an i7 w/ turbo boost. It was not my experience. It’s crazy how fast you can get hot at turbo boost speeds. Under 1ms to go from 40c to 100c… a heatsink can help that a lot, but your conditions have to be pretty ideal for it to work… and that’s nothing to say about a DIY setup where the bond between the CPU and heatsink could possibly be less than ideal.

This is red flag for sure. That heat has to go somewhere. For a heatsink to work, the air has to get hot enough to “rise” quickly and have the cold air replace it from below via convection. If your room around the device (not around your body) is too hot, convection efficiency is reduced. If your chassis isn’t warm/hot, then the CPU isn’t generating heat (unlikely with Roon), or that heat is being held in the air within the chassis.

We’ve see people throw a Nucleus in a closed drawer and then complain that it shuts down. Well, yah… eventually that drawer becomes not enough to cause convection to slow to the point where it just isn’t efficient enough.

I’m not saying you have any of these problems, but you do have a lot of issues with ROCK that thousands of others don’t. There might be something wrong.

Things to consider:

  • Cirrus7 are testing in a 21C room. Unsure how their device would be impacted by a warmer room.
  • Can you reproduce their results on your machine using their test methods?

That would depend on how one was using Roon, wouldn’t it? FWIW, I use an Intel Core i5-8600K on a mini-ITX board in a fanless HDPlex case. I don’t do high res, nor MQA or DSD, and don’t use DSP, either. Just streaming local files or TIDAL. On Ubuntu server I observe load average of 0.08. I live in tropical country and in the mornings, before air condition is a must, the temperature in my room rises regularly above 34C. I have never seen my CPU temperature rise above 45C. My chassis is lukewarm, too. And still I notice the temp difference between the right side with heat pipes vs the left side without them. Only @pl_svn would know how he uses Roon, but without DSP and heavy signal processing, the processor use IMHO is not enough to overheat a CPU in a passive-cooled custom-built chassis.

Danny… it’s been running flawless since last August when I bought it

oh… wait: it did shut down once, at night, one/two weeks later, maybe because of the 12v/5A LPS I was using
bought a 19v/5A one and… never had to reboot since: just Roon Core (auto)restarts after an update

only during these two last Roon updates something went wrong, as I wrote

oh, and… room temperature has been everywhere between 19/20 and 30/35 since I have it. it’s around 24/25 max these days

oh, and 2… the Qobuz favourite albums issue was probably just a temporary Qobuz glitch: synch was instantaneous today :wink:

no DSP here apart x2 upsampling to a filterless NOS DAC (Lumin U1 Mini in-between)
processing speed still over x30 for 192kHz files
no DSD

and even if… glitches only arose during a Roon update and only affected the update itself
all running fine as always afterwards

updates cause CPU spikes when things in the database change, and they often force metadata updates.

My CPU spikes to 100% pretty much every time I get a metadata update, or do analysis of new content. There are many other cases where it spikes as well.

that’s after the update is applied! and… not the faintest issue there

again: issue with latest update is that Roon did not restart
ROCK was still running

NUC was sitting idle since the day before (update applied in the morning)

I love your tactic here. Let’s blame problems that the poster is having on the fanless case so we can deflect from the idea something might actually be wrong with Roon…Heaven forbid!

How about asking questions that might help isolate what is going on with Roon?

I’m not sure what makes you repeatedly put your foot in your mouth when you have nothing valuable to contribute, but let me explain my thinking here:

@pl_svn has had ongoing issues for a while, thus his repeated request for an auto-reboot schedule. I’m addressing the larger issue with a new idea since no one has any idea what’s going on with his install. It was just an idea, and I stated that it may not be applicable, but worth a shot.

As for the update issue, we have ~71,000 installs that have updated to 1.7.555 in the last 5 days. When a handful of people (or less) complain about something, either it’s the start of a widespread problem, or its environmental issue limited to those reporting the problem. 71,000 is not the majority yet (it’s only been 5 days), but we’ve heard or seen nothing pointing to a widespread issue here.

So, nothing I can really do about @pl_svn’s issue, especially since he’s running an unsupported configuration. The symptoms of “stuck on shutting down” point my senses to a hardware issue, but nothing obvious for a stock NUC8i7BEH. Only when @greggl commented on the Cirrus7, did I make the connection and wondered out loud.

I have no answer, I could not think of any extra information to ask about at that time, and the setup is unsupported. If he was running Ubuntu or some other Linux distribution on that system, there is much I can ask about, but nothing with ROCK.


2 questions:

  1. What is your update setting? Do you have it auto-download?
  2. Is it possible you skipped an update?