Local display function fails with HDMI in moOde 7.0.1

[nosferatu_cat]

I've seen this behavior, more often with local display but also using the browser interface. I don't know if the cause is the same as the one under discussion or not but for me symptomatic relief always comes from doing a fresh install. It usually begins with slow refresh when playing music--the current file on the display is not the one playing, and eventually degenerates into total lack of response to input events. It began for me in one of the v 6 revs, can't remember which one. I wish I could be more specific. The intermittent problem is the worst to diagnose, especially when the symptoms are vague. That's when I throw up my hands and fall back on memory leak, disk corruption, or sun spots.

[TheOldPresbyope]

@Alaini93 

Just a note to say that Tim and I have run enough tests to convince ourselves this is likely a bug in the version of Chromium included in Raspberry Pi OS 10.6. Not something we can fix.

One thing you could try is running your moOde player with an Ethernet connection and with the WiFi adapter turned off. In limited testing, it appears that in this configuration the system is able to recover from the runaway memory consumption instead of going catatonic. Mind you, this strategy hasn't been extensively tested so it could turn out not to be useful in real-world usage.

@nosferatu_cat 

It's not clear you are describing the same issue. Better to start a new thread which can be cross-referenced to existing threads if necessary.

Regards,
Kent

[Alaini93]

Just a note to say that Tim and I have run enough tests to convince ourselves this is likely a bug in the version of Chromium included in Raspberry Pi OS 10.6. Not something we can fix.

Darn, so that's mean I need to go on the Raspberry forum and bitch about it?

One thing you could try is running your moOde player with an Ethernet connection and with the WiFi adapter turned off.

I' using Ethernet connection and with WiFi turned off.

Thanks, Like I said, I'll stay with 671 for now.

[TheOldPresbyope]

@Alaini93 

Darn, so that's mean I need to go on the Raspberry forum and bitch about it?

No.

[TheOldPresbyope]

@Alaini93 
@Tim Curtis 

I was reading the release notes for the 2021-01-11 release of RaspiOS and noticed this:

 * Chromium version 86.0.4240.197 included

It turns out this same version is now in the testing repo for the release of RaspiOS on which moOde 7.0.1 is built.

When I installed this version of Chromium to a moOde 7.0.1 player (on an RPi3B+ with local HDMI display) and reran my monitoring test overnight I did not see the failure mode we've discussed previously.

I'm curious to know if you experience the same result.

Code:

# first, remove old version
sudo apt-get remove chromium-browser

# then install the testing version
sudo apt-get install -y -t testing chromium-browser

# reboot if you're as superstitious as I am.

There's no need to make any other changes to moOde (notably, not to /home/pi/.xinitrc)

Regards,
Kent

[Tim Curtis]

Interesting. I can add to the TODO list to bump to the testing version for 7.1.0

[TheOldPresbyope]

@Dmcgee 

Sorry, I missed your response at the time and finally read it today.

I know the kind of keep-alive hackery you mention. I've been on both ends of it---implementing a bodge in desperation to make something work and trying to figure out the bodge someone else slipped in without documenting it

Yeah, my experiences might have involved "a large maIn frBme coMpany" but then again it might have involved another of the nine big computer companies of the 1960s-1970s. I wish I'd kept my pocket instruction cards for all the big iron I burned my fingers on. Fortunately, there's on-line sources of scans of the technical docs for many that I can peruse for a trip down memory lane.


It appears we won't have to resort to such things if this recent version of the chromium browser clears up the underlying problem.



Regards,
Kent

[Alaini93]

Good morning Kent,

I think you fixed it. 30 mins and still working, I usually had problem after 20 mins.
let's see after few hours.
Thank you.
Alain

[Alaini93]

I talked too soon

[TheOldPresbyope]

I talked too soon

Sigh. I'll recheck my test results from last night but I'm reasonably sure I reproduced the same conditions which caused my test to fail in the past.

However, the failure mode I observed would occur after idling for hours. Neither Tim nor I were ever able to reproduce your "in 20 minutes" finding. I don't think we know yet what is different in your case.

Regards,
Kent