apport looping on an old crash dump

Bug #1516947 reported by Alan Pope 🍺🐧🐱 🦄
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
apport (Ubuntu)
Fix Released
Medium
Martin Pitt

Bug Description

A few occasions now I've had apport spinning a CPU on an old caja crash found in /var/crash.

Talked to pitti in #ubuntu-devel who asked for the crash file, which I've uploaded to:- https://chinstrap.canonical.com/~alan/_usr_bin_caja.1000.crash

Some output from stracing the apport process:-

http://paste.ubuntu.com/13310423/

ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: apport 2.19.2-0ubuntu6
ProcVersionSignature: Ubuntu 4.2.0-16.19-generic 4.2.3
Uname: Linux 4.2.0-16-generic x86_64
ApportLog:

ApportVersion: 2.19.2-0ubuntu6
Architecture: amd64
CurrentDesktop: MATE
Date: Tue Nov 17 09:44:05 2015
InstallationDate: Installed on 2014-06-16 (519 days ago)
InstallationMedia: Ubuntu 14.04 LTS "Trusty Tahr" - Release amd64 (20140417)
PackageArchitecture: all
SourcePackage: apport
UpgradeStatus: Upgraded to xenial on 2015-11-03 (13 days ago)

Revision history for this message
Alan Pope 🍺🐧🐱 🦄 (popey) wrote :
description: updated
Revision history for this message
Martin Pitt (pitti) wrote :

According to Alan's strace this hangs in/loops over reading the existing .crash file; this can only be the reading ofCrashCounter.

This reproduces easily with:

python3 -c 'import apport.fileutils; print(apport.fileutils.get_recent_crashes(open("_usr_bin_caja.1000.crash", "rb")))'

Changed in apport (Ubuntu):
status: New → Triaged
assignee: nobody → Martin Pitt (pitti)
importance: Undecided → Medium
tags: added: rls-x-incoming
Revision history for this message
Martin Pitt (pitti) wrote :

Fixed in trunk.

Changed in apport (Ubuntu):
status: Triaged → Fix Committed
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package apport - 2.19.3-0ubuntu1

---------------
apport (2.19.3-0ubuntu1) xenial; urgency=medium

  * New upstream microrelease:
    - apt/dpkg: Fix source record lookup in install_packages. Thanks Brian
      Murray!
    - hookutils.py, attach_gsettings_schema(): Don't replace the schema
      variable; fixes attaching relocatable schemas. Thanks Sébastien Bacher!
    - generic hook: Limit JournalErrors to the 1.000 last lines. This avoids
      long report load times when processes cause massive log spew.
      (LP: #1516947)
    - Add key filtering to ProblemReport.load().
    - Don't read the entire report when determining the CrashCounter. This
      avoids long delays for existing large reports.
    - test_python_crashes.py: Be less sensitive to the precise names of
      gvfs-metadata D-Bus service files.
    - Move backend_apt_dpkg -dbgsym test cases to Ubuntu 15.10.
    - Tests: Move to unittest's builtin "mock" module.

 -- Martin Pitt <email address hidden> Tue, 08 Dec 2015 09:49:00 +0100

Changed in apport (Ubuntu):
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.