Sandbox Debugging

The macOS sandbox confines Chrome processes that handle untrustworthy data or code (or both). It works by blocking the process from accessing OS resources, which can have unintended side effects that cause compatibility issues or crashes. This document provides instructions for debugging such issues.

Determining if the Sandbox is Responsible

The easiest way to test if the sandbox is causing a compatibility issue is to temporarily disable it. This can be done like so:

open -a 'Google Chrome' --args --no-sandbox

Running with --no-sandbox is an insecure configuration. After performing the test, you should quit Chrome and re-launch normally.

If the issue still persists, then the issue is not caused by the sandbox. If the issue no longer occurs, continue to help provide further debugging information.

Debugging the Sandbox

If you have determined that the sandbox is responsible for the issue, the next step is to determine what the sandbox is blocking. Sandbox violations are written to the system log (rather than Chrome's standard output/error), so a separate command is needed to access the data.

Start the log command to show sandbox violation errors:

log stream --predicate '((processID == 0) AND (senderImagePath CONTAINS "/Sandbox")) OR (subsystem == "")'

Then launch Chrome with with the --enable-sandbox-logging argument:

open -a 'Google Chrome' --args --enable-sandbox-logging

After you reproduce the issue, quit Chrome and Ctrl-C the log command to stop it. Copy and paste the entire output of the log command to a text file and attach it to the bug tracker.

You can also access historical sandbox violations using the log command like so:

log show --start '2020-09-21 17:45:00' --predicate '((processID == 0) AND (senderImagePath CONTAINS "/Sandbox")) OR (subsystem == "")'

Adjust the --start (and potentially add an --end date/time in the same format) to limit the amount of output.

Breaking on Sandbox Violations

In order to determine what is causing a sandbox violation, it can be helpful to use the send-signal action on a sandbox rule, like so:

(deny file-write (path "/foo/bar") (with send-signal SIGSTOP))

That will cause the process to stop until a debugger is attached or SIGCONT is sent to the process.