Unrecoverable failure in required component org.gnome.Shell.desktop

[Camille Rodriguez]

A Lenovo P53 is installed with ubuntu 18.04.5 LTS.
Graphics Intel UHD Graphics 630 (main display), nvidia quadro RTX 5000 also installed but not used for main display.

Laptop is domain joined. When logging to the local ubuntu user, there is no problem detected. Then, if we log out and try to login with a domain user, gnome crashes.

From what I can see in the logs, the user is authenticated correctly, but then in the instantiation of the session, gnome.shell.desktop faces an unrecoverable failure.

From auth.log

```
pam_sss(gdm-password:auth): authentication success; logname= uid=0 euid=0 tty=/dev/tty1 ruser= rhost= <email address hidden>
Oct 23 16:27:28 WLLNXPFXXXXX gdm-password]: pam_unix(gdm-password:session): session opened for user <email address hidden> by (uid=0)
Oct 23 16:27:28 WLLNXPFXXXXX systemd-logind[1200]: New session 17 of user <email address hidden>.
```

From syslog
```
Oct 23 10:17:42 WLLNXPFXXXXX gnome-session[1914]: gnome-session-binary[1914]: WARNING: Application 'org.gnome.Shell.desktop' failed to register before timeout
Oct 23 10:17:42 WLLNXPFXXXXX gnome-session-binary[1914]: WARNING: Application 'org.gnome.Shell.desktop' failed to register before timeout
Oct 23 10:17:42 WLLNXPFXXXXX gnome-session-binary[1914]: Unrecoverable failure in required component org.gnome.Shell.desktop
Oct 23 10:17:42 WLLNXPFXXXXX gnome-session[1914]: gnome-session-binary[1914]: CRITICAL: We failed, but the fail whale is dead. Sorry....
Oct 23 10:17:42 WLLNXPFXXXXX gnome-session-binary[1914]: CRITICAL: We failed, but the fail whale is dead. Sorry....
Oct 23 10:17:42 WLLNXPFXXXXX at-spi-bus-launcher[2026]: XIO: fatal IO error 11 (Resource temporarily unavailable) on X server ":0"
Oct 23 10:17:42 WLLNXPFXXXXX at-spi-bus-launcher[2026]: after 21 requests (21 known processed) with 0 events remaining.
```

I will include the full syslog and auth.log to this bug. Please let me know if any other log could be useful to troubleshoot.

[Sebastien Bacher (seb128)]

The log has error about

acvpnagent[1117]: Function: OnOpenRequestComplete File: ../../vpn/Common/IP/HttpProb
eAsync.cpp Line: 313 Invoked Function: CHttpSessionAsync::OnOpenRequestComplete Return Code: -31588316 (0xFE1E0024
) Description: SOCKETTRANSPORT_ERROR_CONNECT_TIMEOUT

it could be the problem leading to hit the timeout on the shell component start

[Camille Rodriguez (camille.rodriguez)]

The vpn client is not causing the timeout. However, I found a lot of kerberos errors in the auth.log and syslog upon subsequent attempts at solving this issue.

(Tue Oct 27 10:54:53 2020) [[sssd[ldap_child[9515]]]] [ldap_child_get_tgt_sync] (0x0010): Failed to init credentials: Preauthentication failed
(Tue Oct 27 10:54:53 2020) [[sssd[ldap_child[9516]]]] [ldap_child_get_tgt_sync] (0x0010): Failed to init credentials: Client '<email address hidden>' not found in Kerberos database

I'm trying to figure out the correct configuration for sssd + realmd + kerberos to be able to join my computer to the AD domain.

[Heather Ellsworth (hellsworth)]

Well From the auth.log I see this:
Oct 23 16:53:18 WLLNXPF1ZY3PG realmd[4214]: Resolving wks.cie.com failed: No DNS record of the requested type for “_kerberos._udp.wks.cie.com”
Oct 23 16:53:18 WLLNXPF1ZY3PG realmd[4214]: Error resolving “wks.cie.com”: Name or service not known
Oct 23 16:53:18 WLLNXPF1ZY3PG realmd[4214]: * No results: wks.cie.com
Oct 23 16:53:18 WLLNXPF1ZY3PG realmd[4214]: * No results: wks.cie.com

And then from the same corresponding timestamp in the syslog we see the NXDOMAIN that was received.

Based on the error in auth.log, could this be an issue with encoding? I wonder where the extra characters are coming from in “wks.cie.com”

[Sebastien Bacher (seb128)]

Hum, which viewer did you use?

Opening the log in gedit it displays
'Resolving wks.cie.com failed: No DNS record of the requested type for “_kerberos._udp.wks.cie.com”'

the string uses utf quotes

[Dean Henrichsmeyer (dean)]

Unsubscribing field high as desktop isn't covered under the Field SLA.

[Camille Rodriguez (camille.rodriguez)]

The problem was more with sssd than gnome.

Not sure of the exact source of the problem with SSSD, but applying the following worked as a workaround :

- Run `sudo systemctl edit sssd.service` and add the following:
[Service]
RuntimeMaxSec=1800
Restart=on-failure

- Restart SSSD:
$ sudo systemctl restart sssd.service

This will restart sssd.service every 30 minutes unless you manually stop it, in which case it'll stay stopped until you manually start it again.

[Camille Rodriguez (camille.rodriguez)]

Re-Actually , we kept getting the failure to login and the gnome crash even by restarting sssd regularly.

I finally isolated my problem to be a duplicate of this bug https://bugs.launchpad.net/ubuntu/+source/at-spi2-core/+bug/1775392

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.