BackGround audio analysis speed stalls [Investigation Started]

Hi @Mark_Sealey,

I wanted to make sure we got you some more detailed info before the weekend hits! Our dev team took a closer look into things, and saw a handful of I/O based errors in relation to a USB drive in your system, that may indicate a potential hardware failure, leading to the analysis hiccup. More specifically:

usb 2-5: USB disconnect, device number 2
blk_update_request: I/O error, dev sdb

Are you able to check the health of this drive?

We’ll be on standby for your update :pray:

1 Like

Thanks, @benjamin!

Appreciated :slight_smile: .

I do have several USB drives - for backups/clones.

(I’m not getting any errors experienced or reported, though. In fact I run Carbon Copy Cloner frequently with all my USB drives as targets; I believe its safety check routines would detect a faulty drive.)

But of course I must look into this with you. Thanks.

OK. Pursuing that… but - just for me to be clear, please - how would another drive unassociated with my database on my Nucleus - interfere with a process running on my iMac?

I hope so!

Would you kindly advise how I can identify which drive this may be, please?

I’ve looked at all the devices on the USB buses using System Information > Hardware > USB.

But there’s nothing that stands out as corresponding with:

‘usb 2-5’ or ‘device number 2’

My devices are listed with (variations on) this set of attributes:

Product ID: 0x0508
Vendor ID: 0x1397
Version: 1.12
Speed: Up to 480 Mb/s
Manufacturer: BEHRINGER
Location ID: 0x14310000 / 11
Current Available (mA): 500
Current Required (mA): 500
Extra Operating Current (mA): 0

etc.

I’m comfortable with the Command line etc… macOS 13.6.7; if you’d let me know, please, how I can identify:

usb 2-5

and/or

`device number 2’

I’ll try disconnecting it from the USB device tree.

Thanks…

Hi @Mark_Sealey ,

According to the logs, this USB also has the following identifier, hopefully, this helps narrow it down:
RoonDBExFAT_PATRIOT

1 Like

Thanks, @noris -

Yes, I know exactly what this is.

My main backup of my Roon database (not music files) is on my iMac. (I always perform backups exclusively from Roon’s own Settings > Backup - sometimes more than once in any one day.)

But every 14 days I - additionally - attach this 64 GB thumb drive formatted exFAT directly to my Nucleus and run a separate backup - again only ever from Roon’s Settings > Backup) as described in this thread.

Although my gut somehow tells me this USB ‘clue’ may be a false lead because I’ve experienced this phenomenon - as have several others on the forums (as I know you know :slight_smile: ) - before I bought my Nucleus last November, I’d of course hope you can investigate further.

For instance: is it possible that Background Audio Analysis was being performed at the moment I ejected the exFAT thumb drive - even though I had waited for the Roon backup to finish completely - and that somehow Background Audio Analysis must rely on external backups to complete?

I get no errors and am able to verify the USB drive in question. Also the male-to-female USB 3 Type A connectors etc.

Thanks again for helping this :slight_smile: !

Hi @Mark_Sealey,

Thanks for the additional info. Another follow-up test, what happens if you remove the below file:

/roon/sys/storage/mounts/InternalStorage/Baroque (C17, C18)/Handel Semele/03-22 O Terror and Astonishment (Live).flac
1 Like

Hi @benjamin; thanks for investigating further.

I was very hopeful: you have obviously found something, haven’t you.

This is what I did:

  1. I keep all my sources, so I located my original of this Album
  2. I ran it though Yate again (and made some corrections to metatags); I also used Yate to check for possible corrupt FLASC headers. None
  3. I then listened to 03-22 in QuickTime; it played perfectly
  4. back to Roon, where I deleted the whole of that Album
  5. Settings > Library > Library Maintenance > Clean up library
  6. Quit Roon and re-imported it (including (a ‘revised’) 03-22) onto my Nucleus; this would be around 16:30 Pacific Daylight Time in the US
  7. crossed my fingers. No errors etc

Unfortunately, no go. Although I got no errors, I still get a hang with whatever seems to constitute a set of (final?) four tracks before completion:

But I see that you think that there is one set of four tracks (etc) which is causing the hang.

It’s always been four since we opened this ticket in January. But of course it’s a different set of four each time I import a new Album or handful of albums.

Please continue to look; I’m standing by to repeat the process immediately with any other suspect.

Or - should I actually do what you asked me to do (:smirk:) and remove just that file for some other reason?

I know I can temporarily appear to clear the hang both with Settings > Library > Background Audio Analysis > Off (and back On); and by restarting the Roon server. I did the latter about 30 minutes later.

Thanks again, benjamin and team :slight_smile: !

But if we can find the underlying cause, please, I’d be grateful.

Hi @Mark_Sealey,

They’re still showing as corrupt on our end - could you please test out fully removing the files from your watched folder within Roon and see if the same issue occurs?

Thanks, Benjamin!

Yes. Starting around 16:15 PDT in the US (California) today, I:

  1. deleted all files for Baroque (C17, C18)/Handel Semele/
  2. Settings > Library > Library Maintenance > Clean up library; 0,0,0
  3. imported a new box set containing 445 files: Couperin Edition - in order to initiate an Audio Background scan; fortunately this also belongs in Baroque (C17, C18)
  4. Background audio analysis began; it detected 445 new files - as it should have done
  5. the process completed at 445 :slight_smile: .

Wonderful.

Presumably I could do one of three things now, @benjamin:

  1. cut my losses and redownload and re-import the whole Semele Album - if we’re sure that this has been causing these hangs?
  2. try importing (all of) /roon/sys/storage/mounts/InternalStorage/Baroque (C17, C18)/Handel Semele/03-22 O Terror and Astonishment (Live).flac to see whether this process has forced Roon to fix things
  3. try importing one track at a time (or some sort of binary chop) of the Semele album to see if I can find a corrupted track… maybe you could help, please, @benjamin, by advising me whether it was only that one (03-22), and I could reach out to the vendor?

Thanks again. This looks like progress, doesn’t it?

Hey @Mark_Sealey,

I’m glad to hear the analysis finished properly!

I believe this may be the best course of action, just to be safe.

Thanks, @benjamin; I have reached out to the vendor…

Have also messaged you with specific questions. I appreciate you are busy and that it may take some time to research answers. So doubly grateful :slight_smile: !