QA:Testcase ABRT vmcore

From FedoraProject

(Difference between revisions)
Jump to: navigation, search
 
(4 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 +
{{admon/warning|kdump in Fedora 19|It appears that kdump does not work properly in F19. See ''Test notification only'' to test at least part of ABRT vmcore functionality}}
 +
 
{{QA/Test_Case
 
{{QA/Test_Case
 
|description=This test case tests the functionality of the ABRT vmcore feature.
 
|description=This test case tests the functionality of the ABRT vmcore feature.
Line 9: Line 11:
 
# Ensure that the system log watcher service is running - {{command|systemctl status abrt-vmcore.service}}
 
# Ensure that the system log watcher service is running - {{command|systemctl status abrt-vmcore.service}}
 
#* If you have to change anything, restart abrtd: {{command|su -c 'systemctl restart abrtd.service'}}
 
#* If you have to change anything, restart abrtd: {{command|su -c 'systemctl restart abrtd.service'}}
{{admon/warning|System-config-kdump is broken in Fedora 17 - edit /etc/default/grub instead}}
+
# Add <code>crashkernel=128M</code> to kernel command line.
# Add "crashkernel=128M" to kernel command line (say, via system-config-kdump -> "Enable", "Apply"). Reboot.
+
#* You can either use system-config-kdump, click "Enable" and then "Apply".
 +
#* Or you can manually edit {{filename|/etc/default/grub}} and then re-generate grub configuration file.
 +
# Reboot.
 
# Ensure that the kdump service is running - {{command|systemctl status kdump.service}}. If it doesn't, no vmcore will be saved on crash.
 
# Ensure that the kdump service is running - {{command|systemctl status kdump.service}}. If it doesn't, no vmcore will be saved on crash.
 
# Crash the machine: {{command|sync; echo 1 >/proc/sys/kernel/sysrq; echo c >/proc/sysrq-trigger}}
 
# Crash the machine: {{command|sync; echo 1 >/proc/sys/kernel/sysrq; echo c >/proc/sysrq-trigger}}
# When rebooting after kernel panic ABRT should detect the new vmcore appearing in {{filename|/var/crash}}, create a crash report and notify
+
|results=
you via the notification area
+
# When rebooting after the crash, ABRT should detect the new vmcore appearing in {{filename|/var/crash}}, create a crash report and notify you via the notification area.
## Results of testing on F15: kexec worked, but apparently it tried to boot full-fledged F15 (instead of minimal kernel which saves vmcore). This failed since 128m is too small for that. Had to power-cycle the machine.
+
 
}}
 
}}
 +
== Test notification only ==
 +
It is possible that kdump does not currently work on Fedora 19: https://bugzilla.redhat.com/show_bug.cgi?id=959914. In this case you can at least test the notification feature of ABRT.
 +
# As root, create a directory under {{filename|/var/crash}} with file named {{filename|vmcore}} in it: {{command|mkdir /var/crash/127.0.0.1-test && touch /var/crash/127.0.0.1-test/vmcore}}.
 +
# Reboot.
 +
# ABRT should present a notification as in the real case.
 +
 
[[Category:Package_abrt_test_cases]]
 
[[Category:Package_abrt_test_cases]]

Latest revision as of 09:14, 6 May 2013

Warning (medium size).png
kdump in Fedora 19
It appears that kdump does not work properly in F19. See Test notification only to test at least part of ABRT vmcore functionality


Contents

Description

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


How to test

Note.png
Please note that this test probably can't be done using live cd.
  1. Ensure you have the plugin installed with the following command:
    • su -c 'yum install abrt-addon-vmcore'
  2. Ensure you have necessary packages for producing vmcore and its processing:
    • su -c 'yum install kexec-tools crash system-config-kdump'
  3. Ensure that the system log watcher service is running - systemctl status abrt-vmcore.service
    • If you have to change anything, restart abrtd: su -c 'systemctl restart abrtd.service'
  4. Add crashkernel=128M to kernel command line.
    • You can either use system-config-kdump, click "Enable" and then "Apply".
    • Or you can manually edit /etc/default/grub and then re-generate grub configuration file.
  5. Reboot.
  6. Ensure that the kdump service is running - systemctl status kdump.service. If it doesn't, no vmcore will be saved on crash.
  7. Crash the machine: sync; echo 1 >/proc/sys/kernel/sysrq; echo c >/proc/sysrq-trigger

Expected Results

  1. When rebooting after the crash, ABRT should detect the new vmcore appearing in /var/crash, create a crash report and notify you via the notification area.



Test notification only

It is possible that kdump does not currently work on Fedora 19: https://bugzilla.redhat.com/show_bug.cgi?id=959914. In this case you can at least test the notification feature of ABRT.

  1. As root, create a directory under /var/crash with file named vmcore in it: mkdir /var/crash/127.0.0.1-test && touch /var/crash/127.0.0.1-test/vmcore.
  2. Reboot.
  3. ABRT should present a notification as in the real case.