Red Hat Bugzilla – Bug 160337
Keyboard "jammed" during smp runlevel 5 boot on IBM HS20-8843 BladeServer
Last modified: 2007-11-30 17:07:07 EST
Description of problem: This problem is EXACTLY the same as Bugzillia report 113925, except it is occurring on IBM BladeServer HS20-8843 hardware. This hardware BIOS does NOT support the solution provided to close bug 113925. I will need a module update that fixes this without disabling USB Legacy Support. Here are the errors from /var/log/messages: Jun 3 10:21:02 sbknpwaq3 gdm[4217]: gdm_slave_xioerror_handler: Fatal X error - Restarting :0 Jun 3 10:21:02 sbknpwaq3 modprobe: modprobe: Can't locate module char-major- 10-134 Jun 3 10:21:06 sbknpwaq3 kernel: mtrr: type mismatch for f0000000,1000000 old: write-back new: write-combining Jun 3 10:21:06 sbknpwaq3 kernel: mtrr: type mismatch for f0000000,1000000 old: write-back new: write-combining Jun 3 10:21:06 sbknpwaq3 kernel: pc_keyb: controller jammed (0xA7). Jun 3 10:21:08 sbknpwaq3 last message repeated 249 times Jun 3 10:21:08 sbknpwaq3 kernel: Keyboard timed out[1] Jun 3 10:21:08 sbknpwaq3 kernel: pc_keyb: controller jammed (0xA7). Jun 3 10:21:08 sbknpwaq3 last message repeated 249 times Jun 3 10:21:08 sbknpwaq3 kernel: Keyboard timed out[1] Jun 3 10:21:16 sbknpwaq3 gdm[4222]: gdm_slave_xioerror_handler: Fatal X error - Restarting :0 Syptom: At end of boot, when system console should be switching from text to graphics mode, the X display never appears, attempt to switch repeats over and over again. Version-Release number of selected component (if applicable): kernel-2.4.21-32.ELsmp How reproducible: Once problem appears, it occurs upon every reboot. Not sure what caused onset of problem. Blades were installed and ran for some days. Next reboot produced problem. It is possible that successful boots were performed on the single-processor grub selection. Steps to Reproduce: 1.Reboot server run level 5 smp. 2. 3. Actual results: Here are the errors from /var/log/messages: Jun 3 10:21:02 sbknpwaq3 gdm[4217]: gdm_slave_xioerror_handler: Fatal X error - Restarting :0 Jun 3 10:21:02 sbknpwaq3 modprobe: modprobe: Can't locate module char-major- 10-134 Jun 3 10:21:06 sbknpwaq3 kernel: mtrr: type mismatch for f0000000,1000000 old: write-back new: write-combining Jun 3 10:21:06 sbknpwaq3 kernel: mtrr: type mismatch for f0000000,1000000 old: write-back new: write-combining Jun 3 10:21:06 sbknpwaq3 kernel: pc_keyb: controller jammed (0xA7). Jun 3 10:21:08 sbknpwaq3 last message repeated 249 times Jun 3 10:21:08 sbknpwaq3 kernel: Keyboard timed out[1] Jun 3 10:21:08 sbknpwaq3 kernel: pc_keyb: controller jammed (0xA7). Jun 3 10:21:08 sbknpwaq3 last message repeated 249 times Jun 3 10:21:08 sbknpwaq3 kernel: Keyboard timed out[1] Jun 3 10:21:16 sbknpwaq3 gdm[4222]: gdm_slave_xioerror_handler: Fatal X error - Restarting :0 Syptom: At end of boot, when system console should be switching from text to graphics mode, the X display never appears, attempt to switch repeats over and over again. Expected results: Gnome Login Screen should appear Additional info: I have been working this problem on RedHat Support Ticket Number 570712.
did some testing to try and replicate this issue, so far the problem has not shown up after a few hundred reboots. Im trying to get more info on this to try and reproduce the issue. Can we find out what type of keyboard was plugged in? normal keyboard port or USB? for point of reference, was there a mouse plugged in as well? PS/2 or USB? any other info on how to reproduce this? does this happen only after gdm errors (seen in the log which was attached)? was the system initially rebooted when X was up and running?
IBM BladeCenter is USB 1.1 internally, so from the OS perspective the keyboard, mouse, floppy and CD-ROM are all USB attached. Externally the keyboard/mouse is PS/2, due to an internal USB->PS/2 converter The BladeCenter holds up to 14 blades, and you only have 1 mediatray (with CD/DVD-ROM, Floppy and USB port) and one KVM. So for 'local' operations, you need to switch KVM and/or MT ownership between them. When you switch KVM and/or MT ownership to another blade, from the OS perspective you actually see USB plug/unplug events. Switching KVM ownership can be done using a button on the front of each blade, using a keyboard combination or using the remote interface. Switching MT ownership can be done using a button on the front of each blade, or using the remote interface. Note: since the JS20 (PPC970) blade does not have a video chip, it cannot be given KVM ownership, and everything needs to be done using SOL (Serial over LAN)
Brian, The Keyboard plugged into the Management Module KVM is a Dell standard keyboard. The Mouse is an IBM PS2 2-button. As Tonko noted above, neither is directly connected to the OS on the Blade. The BladeCenter Management Module handles USB interfaces to each Blade in the rack according to the selected context.
One additional thought on replicating the problem: If you have an IBM HS20- 8843 loaded with RedHat ELsmp 3.0as in an IBM BladeCenter, and repeated reboots do not cause the problem: 1. Make sure you are running the latest available BIOS firmware updates for both the BladeCenter MM and the Blade. 2. Set the Blade OS to boot to runlevel 3 vice 5. 3. Reboot the Blade, then try doing "startx" at the system console, or setting Blade back to runlevel 5 and rebooting.
a patch for this was posted (7/12/05) to rhkernel for approval. the patch was tested and resolves the problem. this issue was due to the fact that the IBM bladecenter blades dont have a console keyboard (hotplug USB keyboard instead). as a result the system is unable to initialize the console keyboard controller/keyboard at boot and spews errors. the fix is simply to skip the console keyboard init if you are on a system without a console keyboard
this is a duplicate of bug 161336 a patch has been posted under bug 161336 that resolves this issue
*** This bug has been marked as a duplicate of 161336 ***
A fix for this problem has just been committed to the RHEL3 U7 patch pool this evening (in kernel version 2.4.21-37.1.EL).
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2006-0144.html