grml-x (via grml-quickconfig) fails to start #20
Comments
|
Installing libgl1-mesa-dri adds 116MB(!) of additional disk space. :( |
|
tried on |
|
That's what I was afraid of, so problems with the |
|
nope. doesn't work. |
|
Oh wow, what does |
|
Xorg Output: Ubuntu Xorg.0.log: |
|
Same on Lenovo Thinkpad T470s (amd64) and EeePC 900A (i386). Might be a general and not a hardware-specific issue. |
|
Installation of xserver-xorg-video-intel package should fix the issue, please confirm whether this fixes your issue, otherwise reopen the bug. Thx! |
|
Reopening the issue, thanks for the update @Head-on-a-Stick - @jkirk do you have any idea about the permission stuff? |
|
JFTR: It didn't help on my Thinkpad T470s either. |
|
@xtaran is there anything in the logs or visible with the tty permissions which might shed light on what's the actual problem? |
|
current daily "auto-build2031 (2017-05-15)" still fails to start Xorg on X250. Xorg.log here: https://gist.github.com/flowolf/56b308dfc4c225edcc0b34e1789734fa can confirm that su'ing to grml user and running startx works (even without xserver-xorg-video-intel) |
|
Thx for all the comments! We have a couple of issues and details here. Please be aware that there are multiple ways to start X but we currently support only a few of them (= these should work). (So please tell us how you start X in your reports. Thx!) We have:
Optionally we have
What (currently) does not work (= is currently not supported):
We should create an issue for each of the supported ways (if they do not work). So to narrow down the problem, this issue should be about "Start x" via |
jkirk
changed the title
|
@Head-on-a-Stick @flowolf @xtaran could you please read the comment by @jkirk and try to reproduce that and let us know what exactly fails and what works? |
When the bootoption "startx" is given, the helper-script "/etc/init.d/startx" is created during the bootprocess (via grml-autoconfig). That means, that if the booptions "startx" is not given, the helper-script is not created and tty6/vt6 should not try to start X but should just start a grml-user-shell. Related to grml/grml#20
When the bootoption "startx" is given, the helper-script "/etc/init.d/startx" is created during the bootprocess (via grml-autoconfig). That means, that if the booptions "startx" is not given, the helper-script is not created and tty6/vt6 should not try to start X but should just start a grml-user-shell. Related to grml/grml#20
grml-runtty used to fix most of the pre-systemd environment problems. It seems that it causes more troubles nowadays than it solved back then. First and foremost grml-runtty does not provide a (systemd) login session which prevents us from starting X reliably. Therefor replacing grml-runtty with agetty with autologin where a grml-shell is used so far. tty1 to tty4, where grml-runtty starts run-welcome (grml-quickconfig), two root- and a grml-GNU-screen session, should also be replaced with a proper systemd-unit. Relates to grml/grml#14 and grml/grml#20
The startx helper script was used to get around some permission problems which were caused by grml-runtty. As we switched to agetty, we also can get rid of this ugly hack. Furthermore we decided to start X on tty7/vt7 as tty6 "is especially reserved for auto-spawned gettys": This is done in order to ensure that there's always a way to get a text login, even if due to fast user switching X took possession of more than 5 VTs. See: http://0pointer.de/blog/projects/serial-console.html Relates to grml/grml#20
The startx helper script was used to get around some permission problems which were caused by grml-runtty. As we switched to agetty, we also can get rid of this ugly hack. Furthermore we decided to start X on tty7/vt7 (if the bootoption startx is given) as tty6 "is especially reserved for auto-spawned gettys": This is done in order to ensure that there's always a way to get a text login, even if due to fast user switching X took possession of more than 5 VTs. See: http://0pointer.de/blog/projects/serial-console.html Relates to grml/grml#20
grml-runtty used to fix most of the pre-systemd environment problems. It seems that it causes more troubles nowadays than it solved back then. First and foremost grml-runtty does not provide a (systemd) login session which prevents us from starting X reliably. Therefor replacing grml-runtty with agetty with autologin where a grml-shell is used so far. tty1 to tty4, where grml-runtty starts run-welcome (grml-quickconfig), two root- and a grml-GNU-screen session, should also be replaced with a proper systemd-unit. Relates to grml/grml#14 and grml/grml#20
To get rid of our ugly hacks to start X via the bootoption "startx", we decided to run grml-x on tty7/vt7 which is defined in the .zprofile-file. When the bootoption "startx" is given, we switch to tty7 (via grml-autoconfig) where agetty automatically starts a grml-user login-(z)-shell and grml-x is started. Closes grml/grml#20
jkirk
changed the title
|
Thank you all for your comments, debugging and logs. They were very helpful. We can say now, that this is not a driver problem, but mostly a permission and systemd issue and I think I have a solution for this. As soon as my PRs got merged I would love to here your experience with following daily image. Thx, guys! |
|
We'd appreciate if anyone of you could verify that it's working as intended, daily ISOs starting with build grml64-full_testing_build2042.iso (https://daily.grml.org/grml64-full_testing/2017-05-25_04-25-15/grml64-full_testing_build2042.iso) are supposed to be working fine. |
Related to grml/grml#44 as well Closes grml/grml#20
|
@Head-on-a-Stick ah, indeed - good catch, thanks for reporting! This should be fixed now with the latest change, I'll verify on my own that this works as intended, though for sure appreciate your feedback too, the daily ISO as of tomorrow (2017-05-29) is supposed to include the fix. Thx! |
|
build 2046 works (Lenovo X250) |
|
Jey, thanks for verification - @flowolf 👍 |
and others
grml-x fails to start on Lenovo X250:
Seems like libgl1-mesa-dri is missing.
The text was updated successfully, but these errors were encountered: