Bug 182623

Summary: 'service NetworkManager status' erroneously reports 'dead but pid file exists' when run without root privileges
Product: [Fedora] Fedora Reporter: Nalin Dahyabhai <nalin>
Component: initscriptsAssignee: Bill Nottingham <notting>
Status: CLOSED RAWHIDE QA Contact: Brock Organ <borgan>
Severity: medium Docs Contact:
Priority: medium    
Version: rawhideCC: caillon, mitr, notting, rvokal
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: 8.32-1 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-04-10 22:56:12 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:
Bug Depends On:    
Bug Blocks: 150221    
Attachments:
Description Flags
Use the pid file in status() none

Description Nalin Dahyabhai 2006-02-23 18:37:50 UTC 
Description of problem:
Running '/sbin/service NetworkManager status' as an unprivileged user reports
that NetworkManager is dead, even if it's alive and well.  I blame this for many
an unnecessary 'sudo /sbin/service NetworkManager restart'.

Version-Release number of selected component (if applicable):
0.5.1-12.cvs20060213

How reproducible:
Always

Steps to Reproduce:
1. Run '/sbin/service NetworkManager status' as an unprivileged user.
  
Actual results:
NetworkManager dead but pid file exists

Expected results:
NetworkManager (pid XXX) is running...

Additional info:
There are at least two things going on here.  The call to the 'status' function
in the init script hasn't been modified to pass the new location of the pidfile
(after bug #179903 was fixed) (quick fix: change 'status $processname' to
'status -p $pidfile $processname').

The 'status' function itself passes the -c flag to 'pidof', which due to a
permissions problem, can't determine which root directory NetworkManager is
using, and omits the right PID from its output.  I don't really know how to get
around this one.  Maybe it should be calling pidofproc() instead of __pids_pidof()?
Comment 1 Dan Williams 2006-02-27 07:18:22 UTC 
Ok, #1 fixed in CVS...

#2 is cute, even the manpage of pidof says "This option does not make sense for
non-root users".

Hopefully this can get pushed to rawhide soon.
Comment 2 Miloslav Trmač 2006-02-27 21:06:09 UTC 
Created attachment 125353 [details]
Use the pid file in status()

Removing the -c can generate false positives in chroots.

This patch is another way to add false positives :) by checking the pid file;
if the daemon is killed and the PID is reused, status () may report a false
positive.  Note that this doesn't help when unprivileged users don't have the
right to read the pid file (e.g. sendmail, {sys,k}logd).
Comment 3 Christopher Aillon 2006-02-27 21:17:54 UTC 
That's a patch to initscripts.  If work needs to happen there, it should go to
notting.