Random SIGSEGVs after porting from Xamarin to net8-android #109459
Comments
|
As mentioned here: dotnet/maui#21479 We found a workaround. The chrashes vanish if we deactivate concurrent garbage collection in the project settings.
I am not sure about the side effects and impacts on performance. When running calculation processes, we see random pauses of the execution for 300ms or more. I assume this happens when the GC is triggered. I haven't found a way to pause the GC for the duration of our calculations. |
dotnet-policy-service
bot
added
the
untriaged
|
Tagging subscribers to this area: @dotnet/area-system-memory |
|
Tagging subscribers to 'arch-android': @vitek-karas, @simonrozsival, @steveisok, @akoeplinger |
|
Tagging subscribers to this area: @BrzVlad |
This issue has been moved from a ticket on Developer Community.
[severity:I'm unable to use this version]
The last major release of our software followed our porting of the software from Xamarin.Android to net8.0-android. Everything appeared to go well in our own tests, but we are getting a lot of reports of crashes out in the wild. We have been unable to reproduce any problem in our office, but with tens of thousands of customers using the software every day, this crashing is a major problem for us.
We've been receiving no managed crash reports via our own telemetry, so we suspected this was a runtime issue. Finally a tech-savvy customer has sent us some debug output, which I'll paste below. We also have some sample stack traces from the Play Store telemetry which I will also paste below.
What I urgently need is for someone knowledgeable on runtime issues to advise if there's a component clearly at fault that we can temporarily turn off or avoid, so we can get back to having a stable app. I could roll back to Xamarin.Android but that is not sustainable long term.
I can provide further log output sent by a customer, but cannot post it publicly due to privacy concerns. Sending it securely to Microsoft will be fine.
Original Comments
Feedback Bot on 2/27/2024, 11:29 PM:
We have directed your feedback to the appropriate engineering team for further evaluation. The team will review the feedback and notify you about the next steps.
Tim Dawson on 2/28/2024, 02:04 AM:
(private comment, text removed)
Tim Dawson on 2/29/2024, 07:52 AM:
(private comment, text removed)
Tim Dawson on 3/5/2024, 07:37 AM:
(private comment, text removed)
Anubhav on 3/9/2024, 10:29 AM:
Tim Dawson Were you able to make any progress in debugging this issue?
Tim Dawson on 3/9/2024, 10:32 AM:
Not really. We can’t reproduce it in the office so we can’t do a lot. I posted it here on the urging of MSFT people on Github, hoping someone would look at the information I’ve made available and shine some light on what might be going wrong.
Timo Pijnappel on 3/18/2024, 02:06 AM:
Any updates on this issue? I am expreriencing very similar crashes in my MAUI app. It crashes during long running processes at random times. Here are some of my logcat outputs:
I have tons of those. All look a bit different. I have no idea where to start searching for a solution. I’m really thankful for any advice.
Jonathan Pryor [MSFT] on 4/11/2024, 11:47 AM:
We would ideally like a repro, or at minimum all the relevant
adb logcatoutput, particularly output before theFatal signal 11 (SIGSEGV), code 1 (SEGV_MAPERR)message. “Relevant output” being output from the originating PID; for example, given this message:we’d want all messages from PID 6482.
For example, consider #96872, which was an assertion failure leading to a crash. This left a message in
adb logcat, but it was before theFatal signal …message:Just the tombstone, or just
adb logcatoutput starting withFatal signal, will not contain such assertion messages.See also #98933, which appears to be the “original” report of the “Random SIGSEGVs after porting” issue.
Tim Dawson on 4/12/2024, 02:41 AM:
Jonathan, did I not upload an entire “tombstone” report sent to us by a customer a few posts up? That contains comprehensive logcat output.
Timo Pijnappel on 8/27/2024, 06:08 AM:
Is there any progress on this? These crashes are still happening, for multiple people as seen in multiple issues on GitHub (Maui, dotnet/runtime). I will gladly assist with debugging. I cannot reproduce the crash, but I can trigger it once or twice a day on my Android device in release mode.
Here is a full logcat of the issue.
LOG-27.08.2024.log
The text was updated successfully, but these errors were encountered: