Red Hat Bugzilla – Bug 200998
"Anaconda" fails to boot graphical installer
Last modified: 2018-04-11 09:32:49 EDT
+++ This bug was initially created as a clone of Bug #200636 +++ Description of problem: After booting from rawhide "boot.iso" as of 2007-07-28 and entering the requested information for an "http" install, "stage2.img" gets downloaded successfully. Hereafter, the installer correctly reports the display type ("LG L1710B" analog) which is connected to the "VGA" output of an "IBM ThinkPad T23". There is a short flicker on the screen as if the "X" server was about to start, buth immediately the screen goes black. The monitor light remains green though which means that at least there is a signal. It is not possible to switch to a virtual console to see what's going on but "ctrl-atl-bckspc" is still able to trigger the reboot of the system. Version-Release number of selected component (if applicable): anaconda-11.1.0.72-1 How reproducible: Always Steps to Reproduce: 1. Boot from corresponding "boot.iso" (2007-07-28). 2. Enter install details and proceed to downloading "stage2.img". Actual results: Screen goes black. Expected results: Graphical installer is launched successfully. Additional info: Adding additional boot parameters such as "noprobe", "skipddc" combined with various resolutions from "640x480" etc. has no effect. Due to an installer regression, current "anaconda-11.1.0.73-1" does not get to the point where the "X" server is started. Hence, I cannot make any statement about that version. -- Additional comment from katzj@redhat.com on 2006-07-31 10:56 EST -- This should be better in trees with .73 -- Additional comment from jfrieben@hotmail.com on 2006-08-01 07:29 EST -- Right, works for me in "anaconda-11.1.0.74-1". I used .74 (earlyer also) from 1 August Build on i386/x86_64 with LG-700E Mointer. but I m facing same problem and GUI is not working, TUI works fine, (after installation GUI failed to start) It is showing Error like: (mini-wm:440): Gtk-WARNING **:cannot open display: : 1 I try with VNC, but same issue on vnc also
This should be better with newer trees
Having closed bug 200636 a week ago, I am puzzled that the graphical installer is broken again: Now (2006-08-05), I also obtain "(mini-wm: 460): Gtk-WARNING **: cannot open display :1" I could swear it worked for me last Friday (2006-08-04). At least in my case, the issue thus is probably related to the the "X" updates of last Saturday (2006-08-05). :(
Created attachment 133711 [details] X.org log from /tmp/ramfs/ directory
It was unable to open display :1 because X had troubles starting up. See the very end of the X.log: Fatal server error: No modes found on either pipe Reassigning to the driver owner.
*** Bug 201639 has been marked as a duplicate of this bug. ***
I addition to comment #2, I have to stress that the issue is by no means restricted to the "i810" driver. Bug 200636 explicitly referred to an "IBM ThinkPad T23" which is powered by a "SuperSavage/IXC 64" chip and which thus uses the "savage" driver. Changing the component to "xorg-x11-drv-i810" does not seem justified to me.
I used two machine for testing one is Intel based (P-IV) and other has ATI (AMD64) Vedio Card.
This only supports my objection to the component change. Maybe Chris somehow overlooked this circumstance?
Graphical FC6 test 2 install on X30 (it has i810 chipset) failed because of 'No modes found on either pipe'. I'll add more messages to other bug reports..
I see Anaconda not trying to use VESA driver like in the past. Anaconda prints out that X server started ok, but the server does not started ("no screens found" at the very end of the log file).
Works for me at least since 2006-08-13 which corresponds to "anaconda-11.1.0.80-1" released on 2006-08-09. You should try a net install of the current development tree.
Same problem with FC6 on "vanilla" desktop PC. Graphical boot and X fail. /tmp/.X0-lock is not being cleared. But, reinstalling grub bootloader will cause graphical boot to work for first one or two times, then it reverts to text mode (most of the time), will occasionally come up in graphical mode. Must "rm -f /tmp/.X0-lock" each time to make startx work.
Does this still occur with the latest bits? There have been lots of changes to the intel driver to detect modes better, so hopefully it works now.
Milan? Could you please confirm whether it is reproducible on the supported distributions?
it was working with 14-JAN-2008 Build anaconda
Thanks for letting us know.