From Fedora Project Wiki

m (replace $ by # because need to be restarted as root)
(updated to match current abrt)
 
(16 intermediate revisions by 5 users not shown)
Line 1: Line 1:
{{QA/Test_Case
{{QA/Test_Case
|description=This test case tests the functionality of the [[Features/ABRT|ABRT]] feature.
|description=This test case tests the functionality of the ABRT kerneloops feature.
|actions=
|actions=
# We will check that the ''dumpoops works and provides correct amount of crashes detected.
# Download files:
#* [http://git.fedorahosted.org/git/abrt.git?p=abrt.git;a=blob_plain;f=examples/oops1.test;hb=HEAD oops1.test]
#* [http://git.fedorahosted.org/git/abrt.git?p=abrt.git;a=blob_plain;f=examples/oops3.test;hb=HEAD oops3.test]
#* [http://git.fedorahosted.org/git/abrt.git?p=abrt.git;a=blob_plain;f=examples/not_oops1.test;hb=HEAD not_oops1.test]
#* [http://git.fedorahosted.org/git/abrt.git?p=abrt.git;a=blob_plain;f=examples/not_oops2.test;hb=HEAD not_oops2.test]
#* [http://git.fedorahosted.org/git/abrt.git?p=abrt.git;a=blob_plain;f=examples/not_oops3.test;hb=HEAD not_oops3.test]
# Test that stand-alone detector tool is working properly:
<pre>
$ dumpoops -s examples_oops1.test
dumpoops: found oopses: 1
Version: 2.6.27.9-159.fc10.i686 #1
BUG: unable to handle kernel NULL pointer dereference at 00000000
...</pre>




 
# Ensure you have will-crash-0.5 installed
# And now test if ABRT's ''KernelOopsScanner'' is working, you'll need root privs for this
#* {{command|su -c 'yum install will-crash-0.5'}}
# Edit /etc/abrt/abrt.conf so it says: <pre>10 = KerneloopsScanner</pre> This makes abrtd scan /var/log/messages more often, every 10 secs. Then restart abrtd:<pre># service abrtd restart</pre>
# Ensure you have the plugin installed with the following command:
# cat the downloaded file to /var/log/messages e.g.:<pre>cat examples_oops1.test >> /var/log/messages</pre>
#* {{command|su -c 'yum install abrt-addon-kerneloops'}}
# Wait for abrt-applet (make sure it's running) to show the notification about the detected kerneloops
# Ensure you have kernel-devel package installed
# Open the abrt-gui and check if you can see the oops
#* {{command|su -c 'yum install kernel-devel-`uname -r`'}}
 
# Ensure that the system log watcher service is running - {{command|systemctl status abrt-oops.service}}
# Make sure {{filename|/etc/libreport/events.d/koops_event.conf}} contains
<pre>EVENT=post-create analyzer=Kerneloops
        # >> instead of > is due to bugzilla.redhat.com/show_bug.cgi?id=854266
        abrt-action-analyze-oops &&
        dmesg >>dmesg &&
        abrt-action-generate-core-backtrace
        abrt-action-save-kernel-data
</pre>
# Run the following command:
#* {{command|su -c 'will_oops'}}
# Observe whether a crash notification appears
|results=
|results=
# After running dumpoops: oops1.test and oops3.test should contain one and three oopses respectively. not_oopsN.test should not be detected as containing oops.
# ABRT should catch kernel oops appearing in {{filename|/var/log/messages}}, create a crash report and notify you via the notification area
# After finishing the second part: oopses detected and visible in abrt-gui
}}
}}
[[Category:ABRT_Test_Cases]]
[[Category:Package_abrt_test_cases]]

Latest revision as of 12:21, 3 May 2013

Description

This test case tests the functionality of the ABRT kerneloops feature.


How to test

  1. Ensure you have will-crash-0.5 installed
    • su -c 'yum install will-crash-0.5'
  2. Ensure you have the plugin installed with the following command:
    • su -c 'yum install abrt-addon-kerneloops'
  3. Ensure you have kernel-devel package installed
    • su -c 'yum install kernel-devel-uname -r'
  4. Ensure that the system log watcher service is running - systemctl status abrt-oops.service
  5. Make sure /etc/libreport/events.d/koops_event.conf contains
EVENT=post-create analyzer=Kerneloops
        # >> instead of > is due to bugzilla.redhat.com/show_bug.cgi?id=854266
        abrt-action-analyze-oops &&
        dmesg >>dmesg &&
        abrt-action-generate-core-backtrace
        abrt-action-save-kernel-data
  1. Run the following command:
    • su -c 'will_oops'
  2. Observe whether a crash notification appears

Expected Results

  1. ABRT should catch kernel oops appearing in /var/log/messages, create a crash report and notify you via the notification area