0

CentOS 7 machine is unable to sign in standard users via GNOME, but root account works just fine.

Steps to recreate the issue.

  1. Select user account and input password and login (user account is part of wheel group)
  2. Background image is displayed but no desktop icons are present
  3. System sits idle for about a minute then goes black and returns to login screen

Journalctl's log file when a standard user signs in is far too large to display here but I noticed I see this particular line a 100 times over. This system is not bound to Active Directory nor does it have any special files added to /etc/pam.d/ or added/modified to /etc/security/.

Mar 07 13:29:03 presstore fsmpm[3391]: pm_query_ldap: got request "uid_by_sid S-1-5-21-552760624-291916025-312552118-255638"

I've captured snapshots of root logging in via journalctl -f and compared them to a standard user and these are the lines that stand out:

Mar 07 13:30:34 presstore gnome-session-binary[6766]: WARNING: Application 'nautilus-classic.desktop' failed to register before timeout
Mar 07 13:30:34 presstore gnome-session[6766]: gnome-session-binary[6766]: WARNING: Application 'nautilus-classic.desktop' failed to register before timeout
Mar 07 13:30:34 presstore gnome-session-binary[6766]: Unrecoverable failure in required component nautilus-classic.desktop
Mar 07 13:30:34 presstore at-spi-bus-laun[6980]: Failed to register client: GDBus.Error:org.gnome.SessionManager.NotInRunning: Unable to register client
Mar 07 13:30:34 presstore at-spi2-registr[6987]: Failed to register client: GDBus.Error:org.gnome.SessionManager.NotInRunning: Unable to register client
Mar 07 13:30:34 presstore at-spi2-registr[6987]: Unable to register client with session manager
Mar 07 13:30:34 presstore gdm-password][5138]: pam_unix(gdm-password:session): session closed for user animation
Mar 07 13:30:34 presstore org.a11y.atspi.Registry[6985]: XIO:  fatal IO error 11 (Resource temporarily unavailable) on X server ":0"
Mar 07 13:30:34 presstore org.a11y.atspi.Registry[6985]: after 21 requests (21 known processed) with 0 events remaining.
Mar 07 13:30:34 presstore com.redhat.imsettings[6776]: [ 1551994234.242747]: IMSettings-Daemon[6860]: INFO: Release the ownership of com.redhat.imsettings
Mar 07 13:30:34 presstore org.gtk.vfs.Daemon[6776]: A connection to the bus can't be made
Mar 07 13:30:34 presstore com.redhat.imsettings[6776]: Exiting...
Mar 07 13:30:34 presstore com.redhat.imsettings[6776]: [ 1551994234.242970]: GLib-GIO[6860]: CRITICAL **: Error while sending AddMatch() message: The connection is closed
Mar 07 13:30:34 presstore com.redhat.imsettings[6776]: [ 1551994234.243036]: GLib-GIO[6860]: CRITICAL **: Error while sending AddMatch() message: The connection is closed
Mar 07 13:30:34 presstore com.redhat.imsettings[6776]: [ 1551994234.243134]: IMSettings-Daemon[6860]: INFO: Unloading imesttings module: gsettings
Mar 07 13:30:34 presstore com.redhat.imsettings[6776]: [ 1551994234.243223]: IMSettings-Daemon[6860]: INFO: imsettings-daemon is shut down.

List of other things I've tried:

  1. Deleted the standard users .config folder
  2. Deleted the standard user and created a new user, unfortunately the problem persisted
  3. Reinstalled gnome packages with yum reinstall gnome-* and restarted gdm

I'm at a loss, I can always recreate the server and install the fiber channel cards, network cards, sas cards, internal raid drivers, and miscellanoes configuration but I'd like to solve the problem or at least understand why it's happening and how to fix it.

I'm thinking I can delete /tmp/.X11-unix and/or /tmp/.ICE-unix but that's just really a shot in the dark.

I believe GNOME might not be loading up because something else is slowing the login process enough that GNOME might just be timing out but I'm not sure. The reason I think so is because before the server would delay a bit then eventually login a user, but now that delay has gotten significantly longer. This might explain why root is able to login because it bypasses whatever check is happening for other users even though standard users are part of the wheel group.

2
  • Is the user's home directory mounted? – Michael Hampton Mar 7 '19 at 23:53
  • Yes it is. I can also login the user via ssh or with other tty consoles. – Aren Tahmasian Mar 7 '19 at 23:57
1

This issue has since been resolved, here's what happened to anyone interested.

A little bit of background.. In our current environment/setup, the SAN volume mount displays its contents with delay for all non root accounts. According to the company that provides the binaries for the SAN Volume to connect to the CentOS machine, the server must be bound to Active Directory in order to remove spurious ldap requests.

Therefore, if a non root user account attempts to get a listing, with either Terminal or GNOME's GUI, a delay of that listing will be induced. Correcting the issue is achieved by binding the machine to Active Directory. Note, the user account can be local to the machine and the delay will still not be present.

Binding the machine to Active Directory resolved the issue and pm_query_ldap logs are no longer present in /var/log/messages or journcalctl.

0

Check to see if you have a file /etc/nologin, also look in /lib, /usr/lib, /run for a similarly named file - I feel I've seen this before I just can't remember the exact filename.

If you find one, remove it and see if that helps.

1
  • Unfortunately I could not find any files named nologin in those directories. I searched the entire root directory just to make sure and all I found was /usr/sbin/nologin. – Aren Tahmasian Mar 8 '19 at 18:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.