Description of problem: Now this worked with fc7 so this is a change problem with fc8 Anyway the system will not start the x server when installing Version-Release number of selected component (if applicable): 11.3.0.50 How reproducible: 100% additional information intel e7221 controler embeded in supremicro mother board Steps to Reproduce: 1. take the rescue cd and boot using the first option 2. wait 3. anaconda says it is running 4. properly identifies the controler 5. Says it is starting the x server. (Doesn't it normally idenfiy the monitor too??? It doesn't in this case 6. then returns with several attempts to start x 7 then returns with can't start x server and wants to know if you want to do text or vnc??? ? Actual results: see above Expected results: ability to install Additional info:
Please attach /tmp/ramfs/X.log and /tmp/XConfig.test to this bug report.
Created attachment 266241 [details] from failed attempt
Created attachment 266251 [details] X config from failed attempt
Have had the same problem with a second similar server so it is probably not a bad hardware issue.
OK I have three machines with similar hardware configurations. All three experience this error. Interestingly the upgrade worked fine otherwise, but in the forced text mode. Then the systems boot up (I always poer boot after a upgrade) and the x server crashes, and again drops you to text mode. But once it does this once, then they start working again. I have looked at all the logs with no indication of why this occurs. But it did occur on all three. Anyway all three are upgraded, so I am happy. But it occurs to me that down the road in the near term someone needs to find out why this happened. The good news is that these are three identical machines for a purpose. One is a production server, which quite frankly it takes an act of God to get permission to take down for even a few minutes. But the other two are test servers. One is a test and standby server for the production server, and the other is a long term advanced test machine for the production server. The point here is that these two machines I can in fact take down without normally requiring outside permission, and so if something additional needs to be tested let me know. I can test it with no problems.
Interesting additional clue. On these same systems, during the boot up process they used to switch to a xwindows monitoring of the boot process. Now they appear to error out when attempting this, and then goes on with the text based boot monitor. Then at the end of the boot process xwindows starts normally and we are off to the races.
I have actaully been seeing this last problem on a lot of machines.
At this point all of my FC8 machines are doing comment 6. I find this very interesting. But otherwise they are working, and generally one should not need to reboot very often.
(EE) GARTInit: Unable to open /dev/agpgart (No such device) Anaconda isn't setting up enough of a /dev for us. Booooo.
Since we've recently removed all the code that has anything to do with setting up /dev, we are most likely no longer failing to set up /dev/agpgart. Is this problem fixed in Rawhide (or F9 Alpha)?
The problem is that both the Rawhide that I can tell and also FC9 Alpha don't support the over the network install, which is the method I use. Any ideas on how I might test this?
I have found out how to do a over the net install of fc9a1 and so far have found that the same machines don't seem to have this problem.
Closing on the basis of comment #12. Feel free to reopen this bug if this problem comes back up.