Open Bug 1643793 Opened 5 years ago Updated 1 year ago

Add diagnostic code for long SQLITE_BUSY retry loops

Categories

(Core :: SQLite and Embedded Database Bindings, enhancement, P3)

Product:
Core
Component:
SQLite and Embedded Database Bindings
Type:
enhancement

Tracking

()

People

(Reporter: haik, Unassigned)

References

Details

We have a couple of high CPU utilization hang bugs[1][2] where we think the cause might be a thread getting stuck in the SQLITE_BUSY handling loop in AsyncExecuteStatements::executeStatement().

This bug is filed to request we add some diagnostic code to help debug this scenario. For example, adding a Nightly-only crash when we are stuck in the SQLITE_BUSY loop for more than 5 seconds. Assuming that would allow us to understand what is causing the SQLITE_BUSY.

  1. Bug 1567018 - Mac Dos/Hang
  2. Bug 1583927 - Firefox has 100% CPU usage

The stack of all the threads and the running SQL (likely sanitized for privacy reasons), it may be enough to figure out something.

Severity: -- → S3
Priority: -- → P3
Product: Toolkit → Core
You need to log in before you can comment on or make changes to this bug.