Open Bug 1722924 Opened 3 years ago Updated 11 days ago

[Meta] Crash in [@ java.lang.OutOfMemoryError]

Categories

(GeckoView :: General, defect, P5)

Unspecified
Android
defect

Tracking

(firefox110 affected, firefox111 affected, firefox112 affected)

Tracking Status
firefox110 --- affected
firefox111 --- affected
firefox112 --- affected

People

(Reporter: fluffyemily, Unassigned)

References

(Depends on 6 open bugs)

Details

(Keywords: crash, meta)

Crash Data

Crash report: https://crash-stats.mozilla.org/report/index/e80a9f00-3fbd-41d4-8f52-652780210729

Java stack trace:

java.lang.OutOfMemoryError

Severity: -- → S2

Agi to create some subtasks that we can schedule in

Flags: needinfo?(agi)
Priority: -- → P3
Summary: Crash in [@ java.lang.OutOfMemoryError] → [Meta] Crash in [@ java.lang.OutOfMemoryError]
Depends on: 1726451
No longer depends on: 1726451

I have been looking at the reports, it looks like a good number of them (about half from what I can see) are while creating an EventDispatcher event, it would be really nice if we could annotate the crash report for those somehow (e.g. with the size of the message and/or the message type).

A suspiciously large number of the crashes is in mozilla::java::GeckoSurfaceTexture::DestroyUnused.

And the third group is from GeckoTelemetryDelegate::DispatchHistogram.

There's a problem with socorro that's preventing me from searching this crashes to get accurate numbers, so I'm gonna wait for that before opening dependents so that I can look at the trends and see if any of these got worse recently.

Depends on: 1726468
Flags: needinfo?(agi)
Depends on: 1726469
Keywords: meta
Depends on: 1796222
Depends on: 1796221
Depends on: 1836432
Crash Signature: [@ java.lang.OutOfMemoryError] → [@ java.lang.OutOfMemoryError] [@ java.lang.OutOfMemoryError: at org.mozilla.gecko.mozglue.GeckoLoader.nativeRun(Native Method) ]
Depends on: 1835126
Depends on: 1854803
Depends on: 1865617

We're seeing a big spike in Java OOM crashes since Fx125 shipped (~10x higher volume than previous releases). Can we please investigate what might have changed to have caused it?

EDIT: Actually, looking at crash-stats, it looks like this might have started spiking in 124.2.0.

Flags: needinfo?(brclark)
Flags: needinfo?(brclark)
You need to log in before you can comment on or make changes to this bug.