Red Hat Bugzilla – Bug 56157
init creates /etc/initrunlvl, doesn't take ownership of it?
Last modified: 2014-03-16 22:24:19 EDT
I just installed SysVinit-2.78-19, replacing SysVinit-2.79-2, which has a
bug which prevents me from using it. I have *not* rebooted since I
installed SysVinit-2.78-19. It appears that the file /etc/initrunlvl was
created as aprt of the installation of SysVinit-2.78-19, but this file
isn't owned by any package.
If SysVinit creates this file, then the SysVinit package should own it.
It's a temporary file. Having these owned causes more problems than it's worth.
(see: numerous posts to enigma-list about 'rpm says this package owns all these
files! I don't have them installed - have I been hacked?'
Besides, the presence of that file indicates something else went wrong when it
was signaling init to re-exec itself.
Well, OK, then that's the bug :-). Why was this file left around when I
installed 2.78-19 on top of 2.79-2?
Not sure off the top of my head. Were there any messages about it failing to
Do you have a /var/run/initlvl?
No /var/run/init* or /var/log/init*. No error messages during the upgrade that
I can recall.
Try it :-).
I did. It worked for me. :)
If you remove the file and run through the steps again, does it happen consistently?
This is what I did to duplicate the problem:
1) Install SysVinit-2.79-2.
3) "rpm -U --force SysVinit-2.78-19.i386.rpm".
After these three steps, /etc/initrunlvl was present.
I was unable to get it to reappear, after removing it, by various combinations
of "rpm -U --force" with either of the two versions of the package. It seems
that the reboot has something to do with the sequence of events.
Perhaps this is something that has been fixed in 2.79, but I'd be hesitent to
close this bug before verifying that. After all, it could be a *new* bug in
I should mention that I'm using kernel 2.2.19, not 2.4.x. I doubt that's
relevant, but it's possible that it is.
Closing out unresolved bugs on older, end-of-lifed releases.
Apologies for any lack of response.