Red Hat Bugzilla – Bug 1278288
unable to start kdump using systemctl
Last modified: 2015-11-06 02:49:16 EST
Created attachment 1089977 [details]
Description of problem:
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.systemctl stop kdump
2.systemctl restart kdump
kdump start successfully
Created attachment 1089978 [details]
Created attachment 1090444 [details]
the log of selinux
In fedora 22, if I
#ls -Z /boot/vmlinuz-4.2.3-200.fc22.x86_64
I will get
But now in fedora 23, if I
#ls -Z /boot/vmlinuz-4.2.3-300.fc23.x86_64
I will get
If I do
## /sbin/restorecon -v /boot/vmlinuz-4.2.5-300.fc23.x86_64
the vmlinuz will change from
then, kdump will get the access on vmlinuz and work.
So, I think this bug maybe caused by the change of selinux, I want to know the difference between
And why selinux change it?
Seems the type of kernel file /boot/vmlinuz-xxx is changed from boot_t to modules_object_t. By executing command "restorecon /boot/vmlinuz-xxx" its type is restored to boot_t, then "systemctl restart kdump" can operate on /boot/vmlinuz-xxx.
Not sure if this is changed on purpose, and is there a way kdump can do to fix it?
*** This bug has been marked as a duplicate of bug 1244057 ***