Red Hat Bugzilla – Bug 189066
system-config-users gives old GID to new users and changes GID of old users with the new ones
Last modified: 2007-11-30 17:11:30 EST
I made a clean install of FC5, and created a user named "josep" with firstboot.
The UID seemed to be 500 and its GID 500. The home directory was created with
the following UID/GID:
$ ls -ldn /home/josep
drwx------ 47 500 500 4096 15 abr 13:20 /home/josep
Also the typical ~/.bash_profile and other files alike (~/.gtkrc) have the same
After a while I created another user with system-config-users, named "mia". When
I look at its ID I guet unexpected results:
$ id mia
uid=501(mia) gid=500(mia) grups=500(mia)
mia's home directory has the following UID/GID:
$ ls -ldn /home/mia
drwx------ 4 501 500 4096 14 abr 13:10 /home/mia
This seems to be wrong, I guess somehow system-config-users used the wrong GID
counter or something.
This wouldn't be a big problem unless that now I have a mix up of files
belonging to different groups. I believe that all files created after the new
user was added have been given the "wrong" group ID.
Looking at /etc/group, it has these two lines (among others):
In the same way, /ets/passwd has the following:
Nedlees to say, I didn't manually tempered with either file, I only used system
If that's relevant, in both files the entry related to "josep" comes always
before the entry related to "mia".
I also created a third user, but it got the correct UID:GID: 503:503, adding a
fourth one did also give correct numbers.
I plan to manually fix this, by editing the appropiate files, and using "chown"
where needed, but I'll wait a while before acting, in case you need more
information from me.
Sorry for getting back on this so late. The current version of the package,
system-config-users-1.2.47, should have some fixes which should avoid that
situation, can you check whether this fixes it for you?
When it came out I installed FC6, and did the same process of creating two new
users, etc, and that problem didn't appear so I guess this is must be fixed now.
For me it's fine to closed the issue now, as I don't think I could help much in
trying to reproduce/debug it, and since I guess I'm the only one that reported
this problem, we can assume it must have been something weird happening on my side.