Bug 145629

Summary:	cups starts on boot after update
Product:	[Fedora] Fedora	Reporter:	Jeff Moe (jebba) <moe>
Component:	cups	Assignee:	Tim Waugh <twaugh>
Status:	CLOSED RAWHIDE	QA Contact:
Severity:	medium	Docs Contact:
Priority:	medium
Version:	3
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2005-01-20 11:37:59 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Jeff Moe (jebba) 2005-01-20 08:40:38 UTC

Description of problem:
After an update of cups it starts on boot, even when the previous
version was set to not start on boot. So if cups is installed, but not
in use, a user who thinks they have it disabled will have it
re-enabled at boot when they upgrade. This may leave them open to
future cups holes (god forbid) when they think the service is off...

Version-Release number of selected component (if applicable):
cups-1.1.22-0.rc1.8.4 and earlier

How reproducible:
Get an older version of cups, turn it off on boot, upgrade.

Steps to Reproduce:
1.
rpm -U --oldpackage cups-1.1.22-0.rc1.8.3.i386.rpm \
cups-libs-1.1.22-0.rc1.8.3.i386.rpm

2.
chkconfig --del cups

3.
rpm -U cups-1.1.22-0.rc1.8.4.i386.rpm \
cups-libs-1.1.22-0.rc1.8.4.i386.rpm

4.
chkconfig --list | grep cups    [shows cups on]


Actual results:
Cups is set to start on boot.

Expected results:
Cups is left at it's previous setting to not start on reboot. Apache,
for example, honors it's previous setting.

Additional info:
I don't quite get why this is happening, but I checked out the
differences between cups & httpd. 

Snippets from their specs:
cups:
/sbin/chkconfig --add cups || true
httpd:
/sbin/chkconfig --add httpd

I don't get why they have different behaviour...

Also, it appears in bugzilla there is an entry for "cup" as well (note
no "s").

Comment 1 Tim Waugh 2005-01-20 11:37:59 UTC

Thanks.  This has been fixed in CVS.

(The "cup" bugzilla component is for the "cup" package.)

Comment 2 Jeff Moe (jebba) 2005-01-22 10:03:02 UTC

Ok. I'm being lazy here as I know I should open separate tickets, but
it's late... These packages have a similar problem:
iiim
pcmcia
sendmail

I'm putting it here so it doesn't get lost before I forget. ;)

Also, what was the voodoo to fix it? It would be handy for us lesser
geeks to know how it was fixed in the ticket.

Thanks for the quick fix!

Comment 3 Jeff Moe (jebba) 2005-01-22 22:08:02 UTC

iiimf: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=145869
pcmcia: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=145871
sendmail: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=145872

Comment 4 Jeff Moe (jebba) 2005-01-23 00:51:57 UTC

Apparently, `chkconfig off` should be used instead of `--del`, so this
is actually NOTABUG (?). Per pcmcia resolution by davej.

Comment 5 Tim Waugh 2005-01-23 09:00:27 UTC

It's not to do with that at all, but caused by the initscript not
being marked noreplace.

Comment 6 Jeff Moe (jebba) 2005-01-23 10:14:01 UTC

Hmm. Ok, then are the others bugs or not? I closed 'em as NOTABUG
after the pcmcia-cs maintainer (davej) said it wasn't a bug. I can
re-open those, leave closed, or forget. ;)

Also, should /etc/rc.d/init.d/cups really be marked noreplace? I mean,
what if there are improvements/fixes to it in future releases?

I think davej may actually be right `man chckonfig`:
--del name
 The  service is removed from chkconfig management, and any sym-
 bolic links in /etc/rc[0-6].d which pertain to it are  removed.
 Note that future package installs for this service may run chk-
 config --add, which will re-add such links. To disable  a  ser-
 vice, run chkconfig name off.

Sorry if I'm being noisy...

Comment 7 Tim Waugh 2005-01-23 15:34:34 UTC

Yes, it should be noreplace.

Dave said that *you* shouldn't use chkconfig --del, and he's right.  The RPM
scriptlet *should* use it, however, since it runs after the package is removed.