Open jsalsburg opened 2 months ago
@jsalsburg, what's in /tmp/capture_RPi_debug.txt
?
This may not be relevant, I have powered down and restarted several times today working on the camera on my workbench.
/tmp/capture_RPi_debug.txt
Mode selection for 2328:1748:12:P SRGGB10_CSI2P,1280x720/0 - Score: 5319.24 SRGGB10_CSI2P,1920x1080/0 - Score: 3319.24 SRGGB10_CSI2P,2328x1748/0 - Score: 1000 SRGGB10_CSI2P,3840x2160/0 - Score: 1648.24 SRGGB10_CSI2P,4656x3496/0 - Score: 2019 Stream configuration adjusted Mode selection for 4656:3496:12:P SRGGB10_CSI2P,1280x720/0 - Score: 13471.2 SRGGB10_CSI2P,1920x1080/0 - Score: 11471.2 SRGGB10_CSI2P,2328x1748/0 - Score: 9152 SRGGB10_CSI2P,3840x2160/0 - Score: 5471.24 SRGGB10_CSI2P,4656x3496/0 - Score: 1000 Stream configuration adjusted Still capture image received
Next time you get the error, look in /tmp/capture_RPi_debug.txt
before restarting Allsky. That file is overwritten every image.
This issue is caused by my Synology which is hosting the AllSky Remote Web Server. When the Synology is transferring files over my Network or is engaged in running its widgets and Applications, apparently the Web Service is not being serviced so the AllSky/RPi cannot push its image files to the Synology Web Server so AllSky times out. I must apologize for reporting this issue before I figured out why it is occurring. It is my fault. As my experience with AllSky increases, I have learned many things, but I am no expert and this issue proves my naivety. This issue necessitates that I create a different Remote Web Server dedicated to AllSky.
@jsalsburg, what model NAS do you have and how old is it? My two-year old, 2-bay, one drive NAS doesn't have these problems. Except when you are viewing a web page on the NAS, the Allsky Website doesn't add any load. And uploading a file every minute should not cause any problems. My NAS can display the CPU time, disk usage, and memory. Perhaps you can view them when your NAS is "too busy"?
No need to apologize for anything - we are both learning.
capture program stopped. This appears to happen right at end of night/begin day. Attached allsky.log. I did not restart the RPi until after I downloaded the debug and log files. Timelapse/Keogram/Startrails files for 20240827 were not generated automatically, however, I was able to generate them with generateForDay.sh. Here is the contents of capture_RPi_debug.txt ... Mode selection for 2328:1748:12:P SRGGB10_CSI2P,1280x720/0 - Score: 5319.24 SRGGB10_CSI2P,1920x1080/0 - Score: 3319.24 SRGGB10_CSI2P,2328x1748/0 - Score: 1000 SRGGB10_CSI2P,3840x2160/0 - Score: 1648.24 SRGGB10_CSI2P,4656x3496/0 - Score: 2019 Stream configuration adjusted Mode selection for 4656:3496:12:P SRGGB10_CSI2P,1280x720/0 - Score: 13471.2 SRGGB10_CSI2P,1920x1080/0 - Score: 11471.2 SRGGB10_CSI2P,2328x1748/0 - Score: 9152 SRGGB10_CSI2P,3840x2160/0 - Score: 5471.24 SRGGB10_CSI2P,4656x3496/0 - Score: 1000 Stream configuration adjusted dmaHeap allocation failure for rpicam-apps0 ERROR: failed to allocate capture buffers for stream
@jsalsburg, please create a new Issue at the rpicam GitHub page and post the Issue number here. What you are seeing isn't an Allsky Issue - it's either a hardware problem or (more likely) something in the libcamera-still application. Let them know that a typical command execution is
libcamera-still --output '/home/rpi/allsky/tmp/image-20240828060449.jpg' --timeout 1 --nopreview --width 4656 --height 3496 --shutter 4077762 --analoggain 1 --quality 100
and that the same command will work sometimes and not other times. If you have changed any Pi settings let them know that as well.
That team has always been very responsive. They will likely ask for additional information. Other people have reported this problem.
Eric
https://github.com/raspberrypi/rpicam-apps/issues/707 dmaHeap allocation failure for rpicam-apps0 # 707
Again, I may be causing this issue. To forgo the limitations of the RPi4 SD card, I attached a USB/SSD to the RPi and altered AllSky to store images to the SSD instead. I may have underestimated the capacity of the SSD and included too many DAYS_TO_KEEP=30, so I changed it to 20 days and deleted 10 days on the SSD. We will see if this solves the issue. This may be a feature to add; a monitor and system alarm/message for /dev/sda* memory value, a display in the image file overlay, something like that generated by the df -P command line.
Other than disk space, I can't think of anything that is impacted by a large DAYS_TO_KEEP. The only thing that looks at the images is the Images
page on the WebUI. I suppose if you had hundreds or thousands of days that page would take slightly longer to list all the days.
This should have no impact on the RPi buffer problem, unless having a USB disk uses a lot of the buffers mentioned in the error message.
Environment
Computer: RPi4 Allsky Version: v2023.05.01_04 Website Version: v2023.05.01 Linux rpi 6.6.31+rpt-rpi-v8 #1 SMP PREEMPT Debian 1:6.6.31-1+rpt1 (2024-05-29) arch64 8 gigs
Bug Description
Every morning, Allsky stops capturing and does not create timelapes/keogram/startrails. I can successfully run generateForDay.sh --timelapse YYYYMMDD etc. If I stop/start Allsky with the 'System' buttons it will stop capture after a few attempts to capture. After rebooting the RPi will continue capture until the next morning when it stops capturing.
Repeats every morning.
Log / configuration files
Cut log file of content before 08-22-2024 to reduce its size so it can be included here allsky.log