X does not load after debian upgrade

Question

After upgrading to "jessie" (debian 8) from stable (debian 7), the system hangs at a blank screen (displaying only an underscore at the top left corner), but otherwise displays no error message to support a diagnosis. The problem seems to be a failure to load X, and the following warnings are contained in /var/log/Xorg.0.log:

[     8.810] (WW) The directory "/usr/share/fonts/X11/cyrillic" does not exist.
[     8.834] (WW) Falling back to old probe method for vesa
[     8.834] (WW) Falling back to old probe method for fbdev
[445797.803] (WW) Open ACPI failed (/var/run/acpid.socket) (Connection refused)
[445800.781] (WW) Open ACPI failed (/var/run/acpid.socket) (Connection refused)
[526625.329] (WW) Open ACPI failed (/var/run/acpid.socket) (Connection refused)
[527190.989] (WW) Open ACPI failed (/var/run/acpid.socket) (Connection refused)

I have tried investigating the error messages relating to vesa, fbdev, and Open ACPI, but haven't found anything particularly helpful. The lack of cyrillic font does not seem very important and ACPI relates to power settings, so I suspect a solution might involve fixing the vesa and fbdev errors. I'm wondering if these messages mean something to anyone, and if so, if a remedy might be suggested.

Edit 1:

Some system information that may be pertinent:

lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09)

uname -a
Linux 3.2.0-4-amd64 #1 SMP Debian 3.2.60-1+deb7u3 x86_64 GNU/Linux

Edit 2:

I was checking an outdated log file (Xorg.0.log). Searching for errors and warnings in the most recently updated log file (Xorg.1.log) reveals the following:

[   232.705] (WW) The directory "/usr/share/fonts/X11/cyrillic" does not exist.
[   232.717] (WW) Falling back to old probe method for modesetting
[   232.717] (WW) Falling back to old probe method for fbdev
[   232.718] (WW) Falling back to old probe method for vesa

Solution 1

From your information, it seems that everything is fine: no errors. But the X display may be on tty8 instead of tty7. So, switch to tty8 with Ctrl-Alt-F8.

EDIT: This is a known problem: https://bugs.freedesktop.org/show_bug.cgi?id=29920

Example of bug where the screen is black while startx works (this may be related or give some ideas): https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=744152

Solution 2

This sounds like your acpid daemon isn't running, for some unknown reason. You can confirm this if you're able to get shell access to the system like so:

$ pgrep -l acpi

You should get some response back. If you do you can double check how acpid is being invoked with this command:

$ ps -eaf | grep acpi

If it isn't running you might want to try forcing it during your system's boot by appending the following to your GRUB line for the kernel. You can do this during the initial boot by appending this:

$ apci=force

These options are discussed further in this article titled: BootOptions from the Ubuntu community wiki.

Comments

[pqnet]

which graphic card do you have?

[vinc17]

The timestamps look strange. Have you rebooted?

[user001]

@pqnet: Sorry to not have added enough pertinent information. I have updated the question accordingly. I notice that uname -a shows deb7u3, whereas "jessie" (debian testing) is deb8. Perhaps the upgrade did not go to completion?

[user001]

@vinc17: I rebooted several times after upgrade, but I'm not quite sure how to interpret the timestamp. I was thinking they might be in ms post-reboot, but no units are provided.

[pqnet]

are you able to switch to a terminal pressing alt+f1 when the error occurs?

[Alen Milakovic]

I'd probably post the whole log somewhere. If it is too long, you could add a link. Also, what is the relevant portion of the x config? Are you using the intel driver? And any third party modules you forgot to upgrade?

[arnefm]

Try switching to another VT after booting (e.g. Ctrl+Alt+F2), logging in and running startx 2>&1 > startx.log and see what turns up in startx.log.

[vinc17]

445797.803 means several days after reboot. So, I don't understand how this is possible. Are you sure that the Xorg.0.log file is the right one?

[Ludwig Schulze]

Can you provide all the latest lines of the Xorg.0.log? tail /var/log/Xorg.0.log.

[user001]

@pqnet: Pressing ctrl+alt+F1 at the frozen screen permits display of the few boot messages that normally occur and then disappear and also shows additional lines Debian GNU/Linux jessie/sid [hostname] tty1. Just below this line is a prompt ([hostname] login:) allowing me to login. Thus, the problem does seem to be X-related.

[vinc17]

Also look at the log messages of your display manager.

[user001]

@arnefm: After doing this, the following line appears in startx.log: /usr/bin/startxfce4: X server already running on display :1. Note that I did this on VT ctrl+alt+F1 rather than ctrl+alt+F2, as in my response to pqnet (though I don't think that this matters).

[vinc17]

display :1 means that the correct X log file should be Xorg.1.log, not Xorg.0.log (which seems to be too old due to its timestamps).

[user001]

@vinc17: Good points. I had just confirmed this by ls -lrt /log/var/Xorg*, which shows log 1 to be the valid file. After searching for warnings and errors in this log file (egrep "WW|EE"), additional warnings were found (added these as an "edit" to the question).

[vinc17]

I have the same warnings. This is not related to the problem.

[user001]

@FaheemMitha: I am using the Intel driver. I don't believe I forgot to upload any third party modules. I'll upload the whole Xorg.1.log in a paste bin. Could you please tell me what you mean by the relevant portion of the x config? Thanks.

[slm]

Are there any msgs in /var/log/syslog? Also take a look at this thread: bbs.archlinux.org/viewtopic.php?pid=825072#p825072

[user001]

@slm: There are 54,000 messages printed to /var/log/syslog, all from today. Thanks for the link. My xorg.conf file is completely empty, so perhaps that is the problem.

[slm]

@user001 - You'll have to comb through those msgs to see what's up, unfortunately 8-(.

[Alen Milakovic]

@user001 The part of the X config concerned with the video driver is what I meant, but you could also paste the whole thing if you want. To be clear, this is usually /etc/X11/xorg.conf.

[user001]

@FaheemMitha: Here are the Xorg log files modified today: pastebin.com/y48ZG9zH

[user001]

@Braiam: Added the full Xorg log files at the link above.

[user001]

Thanks @slm. Seems I was checking the wrong Xorg log file, and the correct log file actually does not contain any acpi warnings (see "Edit 2") in the question. I did try what you suggested and found ktpacpid and ktpacpi_nvramd as active root-owned processes.

[user001]

Thanks. I find that tty1 is command-line only; tty2 is graphical (it's what displays X after running startx via tty1); tty3--tty6 all display a login prompt; tty7 displays a static underscore in the top left corner (which is the same thing I observe when the system hangs on boot); and tty8 displays a blinking underscore in the same position.

[vinc17]

@user001 I suppose that Xorg.1.log corresponds to the server started with startx. I suspect a bug (or misconfiguration) of your display manager. For instance, your display manager may crash before it starts a X server, as startx shows that there are no problems with X.

[user001]

Thanks, I'm looking over the bug report you sent now and trying to see if I can figure out why the display manager is not running on boot.

[vinc17]

@user001 Use bootlogd to get /var/log/boot messages (you should see one or a few lines related to your display manager). If everything is fine here, then you should see where your display manager has sent log messages.

[user001]

During installation, the installer asked me whether I should keep (default action) or replace two existing config files that differed from incoming files. One of the files was /etc/init.d/bootlogd. I opted to retain both of the existing config files (the default action), but since then, nothing is written to /var/log/boot, so I think I should have taken the new config file.

[vinc17]

@user001 Concerning your bootlogd problem, it could be the following bug: bugs.debian.org/cgi-bin/bugreport.cgi?bug=716948

[user001]

I ended up formatting my drive and performing a fresh installation. Evidently, the Debian upgrade process is quite fragile.

[Alen Milakovic]

@user001 YOu are not required to accept an answer if it doesn't fix your problem.