Repeatable B2 Crash

[Nick B-T]

I have a situation where viewing particular tracks via both the browsers pages (Original and new) on different web browsers (Safari and FIrefox) and on the iOS app cause the B2 to behave unpredictably.

When the artist or album are selected from the central column, or when either are returned as the first result of a search (browser or app) the following happen to the B2:

• Goes into “Scanning mode” and returns random numbers for the scan result, all fewer than the number of tracks stored
• Displays “Bluetooth OK”
• Lower three-quarters of the B2 display goes blank
• Red LED flashes rapidly

When I view the artist name on the browser by clicking Rename it displays correctly.  However, when I look at the album details using Rename, the artist is displayed as “Parent artist”. 

My first thought was file or disc corruption (I have a 480G), so I tried the following:

• Scanned - correct results returned.
• Reboot - fine.
• Formatted the HDD (SSD) and reloaded the library
• Started NAS and inspected the file structure (looking from MacOS Finder, not with any tools)

The behaviour continued, and there appeared to be nothing wrong that I could see.

The albums in question, and most others were ripped to FLAC on the Mac, using either gnuDB or Apple Music, the filenames and ID3 data roughly tidied up (why don’t contributors to the databases understand that the composer and the artist are not necessarily the same person, especially for classical music?!?!).  All tracks then loaded via USBA from a LaCie portable HDD from a recommended directory structure (“Music” < *Artist* < *Album* < *Track*)

Only two albums are affected. Formatting and reloading didn’t alter this, so there must be a data problem that isn’t being handled correctly.

When the failure happens, the diagnostic screen spews data to fast to keep up. Here’s the last portion before refreshing the browser:

... (unable to keep up with the diagnostics) ...

...

initDisks () no cookie - waiting for hdd

Trying sda1................. at 358 ms

/dev/b2_internalhdd_p1 partition found

SDA1 found................... at 378 ms

starting initmdb

************ setVolume 25 ************

dacPresent = 0

setAuxLevel ()

setAlsaCaptureLevel () not found - skipping

setAuxLevel () done

setCurrentID (2000000) trackcount = 0

initmdb 1 localPath = /media/hdd1

Current Radio = Jazz FM

setCurrentID (2007543) trackcount = 0

Bass 0 Treble 0

************ setVolume 0 ************

dacPresent = 0

Disk size 480000

calling getUSBSpeed

initmdb () done

First line Version 3

/proc/sys/kernel/core_pattern set to core

Init Disks done at 1188 ms

pic_init () ufd = 9

pic.c: Number of eventX devices changed: 1

initAlsaPlayback: called B2B=1

initAlsaPlayback: called playbackDevice = usbdac

Alsa init complete

playbackThread: Started

ALSA PERIOD TIME 100ms 4410 samples. Buffer size = 22050

ALSA PERIOD TIME 100ms 4410 samples. Buffer size = 22050

Init Play Done at 1375 ms

************ setVolume 0 ************

Mute(oled init)

setBrightness (15)

initVTuner Done at 1612 ms

initBluetooth Done at 1631 ms

testWebThread () started

refreshUI () called by main

1-1.3.1 has manufacturer JMicron

pic_thread (): R1F

fastUnixSocket bind done

usb1 has manufacturer Linux 4.19.23 dwc_otg_hcd

pic_thread (): XE2F455AF50A8E2F2

Waiting for incoming connections on socket

1-1.3.4 has manufacturer JMicron

Security check passed

1-1.3.2 has manufacturer Burr-Brown from TI 

497 albums

7553 tracks

1-1.2 has manufacturer MediaTek

Wifi started at 1778 ms

379 artists

mirrorCount 0

wifiWatchdog starting

Gateway address 192.168.1.254

blacklist id = 0

Sorting artists

************* udpThread started **************

initTV () TV Size 720x480, 32bpp

scanVideos () done 0 videos

New loadPlaylists ()

blacklist id = 0

refreshUI () called by loadOrScan

loadOrScan () finished

Bluetooth adapter path: /org/bluez/hci0

Bluetooth_thread started

Making Bluetooth discoverable

object path /org/bluez/agent registered

Setting Bluetooth Class

Pairing agent registered

Transport thread created

Connecting to socket: /var/run/bluealsa/hci0

No Ethernet Address

Subscribing for events: %b

GOT 0 TRANSPORTS

Bluetooth Init Done

refreshUI () called by bluetooth_thread

a2dp: before main loop

************ setVolume 0 ************

isMinidlnaConfValid()

startSambaIfConfigured ()

Starting Main Loop at 2856 ms

executing /etc/init.d/S91smb restart

initPunchThru ()

lstat result 0 st_mode 41471 S_ISLNK 1

punchThreadCode()

************* sonosDiscoveryThread started **************

Librespot started - PID = 5356

Stopwatch started at 0 ms

Disabling terminal blanking

Done

/etc/inittab size: 1248

noatime ALREADY PATCHED - NO ACTION

readRadiofeeds strings = 1916

freelist malloc = b53e0008

Init Disks at 171 ms

initDisks ()

initDisks () lowSpeed=0

initDisks () no cookie - waiting for hdd

Trying sda1................. at 276 ms

/dev/b2_internalhdd_p1 partition found

SDA1 found................... at 301 ms

starting initmdb

************ setVolume 25 ************

dacPresent = 0

setAuxLevel ()

setAlsaCaptureLevel () not found - skipping

setAuxLevel () done

setCurrentID (2000000) trackcount = 0

initmdb 1 localPath = /media/hdd1

Current Radio = Jazz FM

setCurrentID (2007543) trackcount = 0

Bass 0 Treble 0

************ setVolume 0 ************

dacPresent = 0

Disk size 480000

calling getUSBSpeed

initmdb () done

First line Version 3

/proc/sys/kernel/core_pattern set to core

Init Disks done at 900 ms

pic_init () ufd = 9

pic.c: Number of eventX devices changed: 1

initAlsaPlayback: called B2B=1

initAlsaPlayback: called playbackDevice = usbdac

Alsa init complete

playbackThread: Started

ALSA PERIOD TIME 100ms 4410 samples. Buffer size = 22050

ALSA PERIOD TIME 100ms 4410 samples. Buffer size = 22050

Init Play Done at 1072 ms

************ setVolume 0 ************

Mute(oled init)

setBrightness (15)

initVTuner Done at 1180 ms

initBluetooth Done at 1205 ms

testWebThread () started

fastUnixSocket bind done

pic_thread (): R1F

refreshUI () called by main

1-1.3.1 has manufacturer JMicron

Waiting for incoming connections on socket

usb1 has manufacturer Linux 4.19.23 dwc_otg_hcd

1-1.3.4 has manufacturer JMicron

1-1.3.2 has manufacturer Burr-Brown from TI 

pic_thread (): X9C689F2B14426C58

1-1.2 has manufacturer MediaTek

wifiWatchdog starting

Gateway address 192.168.1.254

Wifi started at 1328 ms

Security check passed

497 albums

7553 tracks

379 artists

mirrorCount 0

blacklist id = 0

Sorting artists

************* udpThread started **************

scanVideos () done 0 videos

New loadPlaylists ()

initTV () TV Size 720x480, 32bpp

blacklist id = 0

refreshUI () called by loadOrScan

loadOrScan () finished

Bluetooth adapter path: /org/bluez/hci0

Bluetooth_thread started

Making Bluetooth discoverable

object path /org/bluez/agent registered

Pairing agent registered

Transport thread created

Setting Bluetooth Class

Connecting to socket: /var/run/bluealsa/hci0

Subscribing for events: %b

No Ethernet Address

GOT 0 TRANSPORTS

Bluetooth Init Done

refreshUI () called by bluetooth_thread

a2dp: before main loop

************ setVolume 0 ************

isMinidlnaConfValid()

Starting Main Loop at 2316 ms

startSambaIfConfigured ()

executing /etc/init.d/S91smb restart

initPunchThru ()

punchThreadCode()

lstat result 0 st_mode 41471 S_ISLNK 1

Librespot started - PID = 5538

************* sonosDiscoveryThread started **************

CD state changed 1

WAKEUP in webListPlaylists ()

Connect index = 1

bluetoothConnect (/org/bluez/hci0/dev_40_EF_4C_C8_93_4F)

BLUETOOTHINITIALISED

No BT MAC address saved

Couldn't Connect: GDBus.Error:org.bluez.Error.Failed: Host is down

Done

refreshSonosActive ()

webGuest url = guest&mode=0&playlist=1&time=1644173910043

setGuestModePlaylist 0

refreshUI () called by webGuest

Clear

[JFBUK]

hi,

Do these two albums playback successfully when you playback the ripped files on your Mac ?

When you tied up the Tags what tool did you use ?

Can you provide the catalog number of one of the troublesome CD's ?

Thanks

John

[PMB]

Hi mbta...,

It sounds like those two Albums are a problem for some reason. Perhaps they have embedded info that the B2 doesn't like. Does the B2 behave OK apart from these albums?

Try deleting them and rip them using the B2.

Paul

Brennan Support.

[PMB]

Hi mbta...,

I also meant to say, it is better to add the debug list as an attachment, saves a lot of scrolling.

Paul

Brennan Support.

[Nick B-T]

Hi, John,

The albums play fine on the Mac as FLAC files. I ripped them with Phile Audio (https://apps.apple.com/gb/app/phile-audio/id413934579?mt=12 )and edited the tags with Squeed (https://www.macupdate.com/app/mac/55832/squeed).  

Details of the album (London Records – 289 452 910-2) here: https://www.discogs.com/release/11861473-Holst-John-Williams-4-Strauss-Zubin-Mehta-Los-Angeles-Philharmonic-Orchestra-The-Planets-Star-Wars-A

The offending track appears to be Also Sprach Zarathustra - Von Der Wissenschaft, where the ID3 track field has the entire ID3 contents dumped into it.  When that is truncated to a sensible length (approx 40 characters) the whole album behaves itself.

... Paul, will add future debug lists as attachments.

Nick

On Friday, February 18, 2022 at 9:55:56 AM UTC JFBUK wrote:

[Nick B-T]

Further to ... I've just looked at the info on GnuDB. It's clear that the info for that track is corrupt. Shows here as item 6 in the list: https://gnudb.org/cd/cle90ec80f 

Having ripped all the CDs on the Mac, and having kept a backup copy, I'm working my way through the entire set, to create normalised tags (my personal normalisation philosophy). Regardless of the source of tags, FreeDB, GnuDB, Apple Music, or anything else, there are, for my liking, too many errors an idiosyncracies in what has been submitted, especially with classical recordings. When complete, I'll reload. Hours of fun!

Paul. I'll rip the album on the Brennan and publish the result.

Nick

[JFBUK]

Hi Nick,

I think you have a reasonable diagnosis on the cause, i.e. the quality of that data in the database referenced by the ripping software.

I'm quite a fan of MusicBrainz Picard software as a tool to look at tag information of a release you have ripped.

That's not just in your ripped file but what is held for the track in the MusicBrainz database.

Again, the quality of the tag information can be dubious as its publicly uploaded.

Picard also has the ability run tag editing scripts but I have not explored that much as yet

https://picard.musicbrainz.org/downloads/

John