-
Notifications
You must be signed in to change notification settings - Fork 707
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Memory leak in latest FW (15.4.0) #2823
Comments
I also confirm. I was wondering since a couple of weeks, why the two ESPs are freezing periodically. After your post I did also turn on the recording of he free memory value and I can see the same behaviour. |
Can someone test the https://github.com/VioletGiraffe/cppcheck-vs-addin plugin to find memory leak ? |
The first part of the graph , before 26th of December, shows exactly this. I was on the previous FW version then (15.3.0), since I always keep it up-to-date. Or are you looking for something else? |
This is exactly the proof that it was introduced in 15.4.0 release. Now need to review the code changes. |
The mains changes are in submodules : code/components/esp-nn and in new MakeStaticResolver() function in code/components/jomjol_tfliteclass/CTfLiteClass.cpp |
Same problem here.. frequent reboot after few hours (due to memory leak) |
I can confirm that there are issues with 15.4! |
I once wrote a memory usage visualizer: https://github.com/caco3/psram-usage-visualizer I will see if i can find something... |
Between 15.3 and 15.4 we updated 3 ESP modules. I tried to revert them, but it does not build anymore. |
Attempt to find leak: #2837 |
@JanHBade thats interesting, the longest mine has been up for is just over 2days. i have 8 digital and 1 analog digits, i wonder if that makes a difference to the reboot? |
The memory leak will be fixed with #2842 You can already use the image from https://github.com/jomjol/AI-on-the-edge-device/actions/runs/7731455511?pr=2842 or wait until it is in rolling or until we created the next release.
It depends on your configuration. I also have one device with only digits where the leak did not occur. |
because it is not in rolling yet! see my post! |
There is a new release 15.5.0 which fixes this. |
It seems to work indeed. Tnx for the quick fix! |
@spanzetta Please stop spreading the wrong thing about the power usage. Your issue is an insufficient power supply! Else I wonder why you are the only one with this issue! |
Removed comment about power supply |
Apparently it's not only me.. |
Please point me to the comment there regarding power issues! |
@spanzetta Hi Stefano, please don´t make your problems with this Open Source SW to ours. You are spreading wrong facts out because your skills unfortunately is not guiding you in the right direction. We are glad to support anyone with issues here, but please don´t go on our nerves. If you have something to report, collect all the facts and report, as it is common in a Bug Fix forum. If you don´t have facts please report to the comment forum, maybe someone can help you there. |
Ok I will stop reporting my experience.. |
@spanzetta |
The Problem
After upgrading to the latest FW (15.4.0) I noticed some readout issues after some time running it. This could possibly be related to #2811.
When I did some digging on my own to investigate what was going on, I noticed that the amount of free memory was stadily decreasing. This can be seen in the attached image, which is a screenshot from my HA history regarding the "Free Memory" propery of the sensor that is exposed of MQTT. Here one can clearly see that the "Free Memory" shows a steady decline over time until it reaches a certain minimum and either crashes and reboots or the FW decides to reboot (unsure if that is implemented; did not dig into this yet).
The FW was installed during the Christmas holidays (the log suggest on the evening of 25th of December around 10:00 PM as that is when this undesired behavior starts). Before the upgrade the free memory was more or less stable over long periods of time.
Version
15.4.0
Logfile
https://www.dropbox.com/scl/fi/2dx9236oqlw3r1d8vjv1g/MemoryIssueLogs.7z?rlkey=z0anabue0rf0fahgvwgh1heno&dl=0
Expected Behavior
"Free Memory" should stay stable over time. This now looks like a memory leak somewhere to me.
Screenshots
Additional Context
No response
The text was updated successfully, but these errors were encountered: