@sdetweil said in Third-Party Module - how to get help?:

https://wpitchoune.net/tricks/raspberry_pi3_increase_swap_size.html

I will give this a shot. Given that each time the slideshow has died, it has indicated that it ran out of memory at roughly 100Mb (and the default swap is 100Mb), this could definitely be related. I’ll try doubling it to see what happens with this module.

I was going to try tackling the transitioning to prevent changing to the new image until it was rendered (to prevent the “painting”), but need the memory issue fixed first. :)

Does your image module contain the JS code you were referring to that hides the image until it’s rendered?

@sdetweil said in Third-Party Module - how to get help?:

app.commandLine.appendSwitch(‘disable-http-cache’);

Thanks. I have made the change and restarted the mirror.

I’m running with only this module and the clock enabled and am finding that having the image caching disabled (so far) is significantly slowing down the rendering of the images on screen.

Is there a way to set the node.js maximum memory or set a limit on cache size? I have found some commentary that seems to say that node.js will use up to 1.7G of RAM by default and have no idea if there is anything in any of these configuration files that would lower that limit using the settings:

--max-old-space-size={limit in KB} --optimize-for-size --max-executable-size={limit in KB}

Given that my system is 1GB of RAM, and the default for node.js is to use up to 1.7GB, I could see where it could conceivably be trying to use memory beyond the bounds of the system, especially if there were some sort of issue with the underlying OS and its interaction with node.js to control the memory usage.

I thought I would follow this up as I have found and solved this specific issue.

Long story short, this particular module has a pre-defined list of file extensions that it will recognize as images. For some reason, I pulled a copy of image files from two different locations and the extension changed from JPG to JPEG. The former is default, the latter is not. I updated the module configuration and it started working again.

I also discovered that the module is running out of memory while running. The error seems to be very consistent although the image in place when it crashes is not. I’ve posted an issue to the module owner’s repository to see what help that might bring.

This line just keeps repeating:

MMM-BackgroundSlideshow received a system notification: CLOCK_SECOND

The only red text is for an error with the weather module then it states that it’s changing to a different forecast.

I’ve been working on setting up the MMM-BackgroundSlideshow module up and running, but am having issues.

I have about a dozen photos that I shot with a 24MP DSLR, edited, and exported as JPG. They are anywhere from about 4MB to about 9MB in size. I’ve copied these into a folder on the Pi and was able to get the module to begin showing them. Then, it would randomly die with an “out of memory” error (from the error log) and the whole screen would go black. Restarting MM would fix it temporarily.

I commented that section of the config out to make some updates to the calendar section. Once done tinkering there, I tried re-enabling the background image stuff and I get zip. No errors, no pictures, nothing.

So, my question is how to go about getting help from the author of the third-party module? How do I determine any details about where the module is dying out and not even loading?

This is a new setup on my first MM, so it may not be related to the calendar at all. In the event that it is, here’s my setup:

{
                        module: "calendar",
                        header: "Shared Calendar",
                        position: "top_left",
                        config: {
                                calendars: [
                                        {
                                                symbol: "calendar-check",
                                                maximumEntries: 25,
                                                wrapEvents: true,
                                                fetchInterval: 900000,
                                                fade: false,
                                                fadePoint: 1,
                                                url: "https://calendar.google.com/calendar/ical/<stripped for security>/basic.ics"                          }
                                                /*url: "webcal://www.calendarlabs.com/ical-calendar/ics/76/US_Holidays.ics"                                     }*/
                                ]
                        }
                },

I have the following issues:

• I do not appear to be getting more than the default ten entries
• I say “appear” because the list of entries is fading at the bottom
• I can’t tell any difference between the default fadePoint and the hard-coded one

I know that these variables are supposed to override the global settings, but I don’t even know if (where?) those are coded if not directly into the code itself.

Does anyone have any thoughts?

@sdetweil said in Initial Setup Q's and Some Doc Feedback:

@ember1205 said in Initial Setup Q's and Some Doc Feedback:

Backing up just a bit, there appears to be no guidance in the install doc as to whether I should choose to allow the installer to set MM to autostart and/or whether I should disable the screensaver during the install process. I chose “N” and “Y” respectively.

that is all really choice on your installation and and usage planning…

i don’t want to touch the mirror… if it reboots cause of power outage, etc, I want it to logon, and start mirror, and never screen lock…

so, set OS to auto logon (manual step)
use pm2 to autostart
and turn off screensaver

Yes. Those were all done. But, the screensaver disable was done during the install and it appears to not have modified the same files as the manual changes show.

@sdetweil

Thanks… Good updates. I will definitely look through your updated installer.

With regard to the warnings - I understand what you’re saying, but my concerns were driven more by the warnings that said “this might not even work.” I realize that they may be very specific to certain aspects, but given that isn’t indicated specifically, it raised up my alert level.

On the fake versus real driver piece, I did notice the “f” to indicate the fake driver. With zero experience with this piece, and having no background that there IS a fake driver or that it’s now default, my eyes told me that the name for the fake driver was actually the name for the real one at first [two] glance[s]. It wasn’t until I reached the latter part talking about the fake driver that I realized it was a different name and changed it.

I’m not using portrait mode as I want this to be a “Magic Photo Frame” and display pictures (many of which are in landscape format), so it seems that I might be able to use either.

Getting started, numerous issues and questions doing the initial setup.

First, I want to say that the overall docs for installing and configuring this appear to be reasonably solid and generally well-maintained - kudos for that!

With that said, there appear to be some gaps that I’m falling into.

The doc I’m using to get started is this one: https://github.com/MichMich/MagicMirror
I have an RPi3 with a 64G card. I am using the full Raspbian image and loaded it via NOOBS. After booting up and stepping through the few minimum steps that the device requires (including setting up WLAN and doing an OS update and then restart), I drop to a shell to begin the MM install and setup. For automatic installation, all I should have to issue (according to the docs) is

bash -c "$(curl -sL https://raw.githubusercontent.com/MichMich/MagicMirror/master/installers/raspberry.sh)"

While that works, there are a LOT of warnings about things potentially not working. So, I opted to update node.js according to the next two steps:

curl -sL https://deb.nodesource.com/setup_10.x | sudo -E bash -
sudo apt install -y nodejs

This clearly processes updates and “makes me feel better.” Was it a necessary step? Should this have been required given that I opted for the “automatic” install?

Backing up just a bit, there appears to be no guidance in the install doc as to whether I should choose to allow the installer to set MM to autostart and/or whether I should disable the screensaver during the install process. I chose “N” and “Y” respectively.

After the node.js updates and a restart, I moved on to the Configuration section and opted to follow this doc: https://github.com/MichMich/MagicMirror/wiki/Configuring-the-Raspberry-Pi

The information in the OpenGL driver section are very confusing. It states that you shouldn’t need to modify the OpenGL driver, then goes on to tell you to edit a file whose location has changed (but the commands still point to the old location - you have to find the info in the notes that the file has been moved), tells you to add a line for

dtoverlay=vc4-kms-v3d

Which looks IDENTICAL to the new default (which is already in the file) as

dtoverlay=vc4-fkms-v3d

That is also referenced as an option later on in the doc, but there’s no commentary that I could find as to whether I should use one or the other for my specific hardware. Also, the file HAS the new default, but it’s in a RPi4 section. Again, very confusing here, and really don’t know what I should have opted for. I chose the full OpenGL and not the “fake” one.

In the Screensaver section, the first portion is already present since I opted to let the installer disable the screensaver. The second portion, however (LightDM) is not only not present but there is no discussion of what this is. Is this actually a required piece? What does it do?

The section for disabling WiFi Power Saving seems pretty broken, at least for my needs. The section about adding information for Pi3 (Jesse and below) will BREAK your setup if using a new release. If nothing else, I would say this entire doc would be immensely easier to follow if there were some direction provided for inspecting the /etc/os-release file to determine which build of Raspbian is being used along with some notes on which names were released in which order (so you know that buster is newer than Jesse, for example).

For the VNC server section, it was quite simple for me to just “turn on” the SSH and VNC server options from the Options panel in the desktop on the RPi when doing initial setup. I also want to point out that doing it this way also requires that the security options on the server be changed from UNIX Password to VNC Password or at least some VNC clients won’t connect.

If anyone has any thoughts on the pieces where I wasn’t sure what to do, I’d appreciate it. Thanks!

@BD0G

Thanks - I’ll keep that module in mind and be sure to load it.

I wonder if the card failure is the result of a disk write operation being in effect at the time of power-off and the level of corruption actually ends up causing the card to be converted into a permanent read-only device (with the corruption). I have seen a SSD drive fail like this, although I don’t believe that it had anything to do with powering off.

At any rate - I’m usually good about having things like this plugged into a UPS and shutting them down properly. :)

@BD0G

Thanks. I have a Pi2 and Pi3, so I can use those. I also have some different cards that I can use for initial testing and such, but I’ll look for something a bit more “durable” in terms of the card itself.

Why would a card fail from a power-off? I could see file corruption, but straight up failure? And, since I don’t actually have adapters with a power button, my choices would only be a cord pull or graceful shutdown. The last time I was testing some functionality for a Pi, I ended up building a simple web system that allowed me to connect from any device and click/tap to initiate a shutdown. I could certainly go that route again if need be.

@sdetweil

Thanks!

Is there a “How to get started” sort of guide somewhere that I should use as a starting point? I see mention specifically of Raspbian, but what other minimum requirements are there? Pi2? Pi3? Pi4? Any recommendations for specific Micro SD cards based on longevity, wear, and price for this effort? Or, should I just go with a DLC type of card if I can find what I need at a reasonable cost?

@sdetweil

Thanks for the comments - good to know that I’m not blazing any trails here! :)

Hi, all.

Long-time linux admin (25 years) and technologist (30 years worth!) that is always looking for ways to add “cool” things to make my home and life nicer and easier. I came across MM recently and am very intrigued in potentially building something out for my home and possibly someone else’s as well.

I have been using an out-of-service iPad Air as a sort of digital photo frame for 9 months or so. While it does show photos, there are some things about it that I would like to change if I could. For example, I have to upload the photos from my Mac into iCloud, wait for them to sync to the cloud, wait for them to be available for the iPad, shut down the slideshow, copy them to the local photo folder, then restart the slideshow.

Similarly, I had a Samsung Galaxy Tab Pro that I set up with Fotoo and gave it access to a specific Google Drive account where I would push photos from my phone (or my Mac). I never had to touch Fotoo, including if the tablet restarted (I configured Fotoo to auto-launch on device startup). It would periodically read the Drive location and pick up new photos and constantly show a randomized photo slideshow. This was very nice for the in-laws while we were on a cruise - we would occasionally upload a couple of photos that we took during the trip, and they would automatically get that photo on their frame.

The Galaxy Tab has been taken out of service due to a battery issue (it swelled and pushed the display out) and I just haven’t gotten around to fixing it yet. The iPad is ok, but at 12’ away the pictures are a bit small. And, as mentioned, I don’t like the way I have to interact with it (including if there’s any notification that pops up on the screen and shuts the slideshow down).

I’d like to take on some testing of MM and would be interested in being able to access a Google Calendar, Weather, and a local file folder on the device to show random photos from. Ideally, I would like to be able to copy down data from a Google Drive location into the local directory so that I’m not constantly pulling the content across my data-capped Internet connection every time I want to show a photo. If that isn’t a viable option, I could use a cron job with rsync to auto update the directory on the local Micro SD card from a Linux server (including across the Internet for the remote device) if I had to.

While I take a lot of photos with my phone and can push them directly to Drive, I also take a LOT of photos with my DSLR when we travel and would like to drop them into a specific folder on my local server after I’ve edited them and have them be shown on the MM/frame.

Is there any sort of module or configuration that would allow me to split a display into multiple areas and have each one show a different photo, and have the photos change on their own random / semi-random intervals?

Any thoughts, help, guidance, etc. would be appreciated. Thanks in advance!