Scientific Linux/RHEL 5.4 04NOV2009 Kernel 2.6.18-164.2.1 on X86_64 Hello! After a successfull installation, I'm unable to boot into the system. Boot process freezes after showing "setting hostname <host>: [OK]". I still get a blinking curses but no reaction to keyboard inputs. I've tried booting without RHGB, using NOUSB, NOACPI and SINGLE and several combinations of the above, all to no avail. I have also disconnected any external hardware and have reinstalled the whole system twice. DVD's have been checked for integrity. How reproducible: On every boot. Even when trying to boot from DVD into rescue mode. Steps to Reproduce: 1. Boot the system Actual results: system freeze/crash Expected results: Should boot to login screen. Additional info: I have been using Debian SID until a few days ago with no problems. (e.g. USB, ACPI, etc. all working perfectly fine) Thanks!
Here are three things that might help us diagnose this: 1. Can you remove 'quiet' from the grub boot menu and see if any more detailed information is available? 2. Try it with 'quiet' removed and 'pci=nomsi' added to the kernel command line and see how it looks. 3. What is the manufacturer and model of this system?
Created attachment 369049 [details] screenshot when it freezes with quiet and rhgb removed
1. I have tried that before, it looks like a normal boot sequence and hangs at the same point. 2. tried pci=nomsi, no change, no new messages 3. Laptop Samsung Q45 Processor: Intel Core 2 Duo T7500 Graphics: Mobile Intel 965 I found out that in some of the older kernels this specific model had major problems with video. So perhaps it has to do with that. I attached a (crappy) picture of what the screen looks like when it dies.
Thanks, Matt. Can you try with 'acpi=off' on the kernel command line as well? If that doesn't work, it might be interesting to try single-user mode by just putting 'single' on the kernel command line as well. Maybe if we can get it booting you can blacklist the video module and then we can try to get it going from there. This could be related, but since that regression was supposedly introduced in 2.6.23, I can't say for sure if it is in RHEL5. http://bugzilla.kernel.org/show_bug.cgi?id=10448
Great job Andy! That fixed it! Using kernel options "pci=nomsi" and "acpi=off" before the crashkernel bit in exactly that sequence, fixes the problem. The Machine now boots as it should and runs perfectly fine. Curiuously, just using either option or both in the wrong sequence doesn't work or even freezes it earlier in the boot process. I'm investigating further... Thanks a lot!!!
Excellent! I found some other reports that this system might have a questionable BIOS that requires booting with the 'acpi=off' option, so if you feel like it, you could always try and update the BIOS.
You won't believe it, I always thought I had kept my system up to date with the latest firmware and software updates. As of yesterday, the nice but pretty useless Samsung update utility that came with the computer, had actually convinced me that the system has the latest BIOS. Well, it turned out I was on version ST7, with the most recent being ST11. Flashed it and voila, everything works with the default kernel options! Once again, many many thanks... (I'm now feeling a bit supid, for not catching this one...hehe) regards, Matt
Matt, I'm glad to hear your system is working well after the BIOS update! Don't worry at all about the time spent on this bug. The great thing about opening this bug and tracking the resolution is now others in the community will find this bug and quickly know that updating the BIOS on the Samsung Q45 to at least ST11 will help them out on almost any modern Linux distro. I'm going to close this as NOTABUG, but please reopen it if you run into any more problems related to booting on RHEL5 or another bug if you've got another problem.