[Bug 723073] New: 12.1, beta1 (x86-64) upgrade from 11.4 hangs on kexec and normal boot at: doing fastboot, creating device nodes with udev
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c0 Summary: 12.1, beta1 (x86-64) upgrade from 11.4 hangs on kexec and normal boot at: doing fastboot, creating device nodes with udev Classification: openSUSE Product: openSUSE 12.1 Version: Beta 1 Platform: x86-64 OS/Version: Other Status: NEW Severity: Major Priority: P5 - None Component: Other AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: abittner@stud.fh-heilbronn.de QAContact: qa@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (Windows NT 5.1) AppleWebKit/535.2 (KHTML, like Gecko) Chrome/15.0.874.83 Safari/535.2 upgraded a normally working clean testinstall of 11.4/x64 to 12.1/beta1/x64 (downloaded iso image file). installing/upgrading packages worked normally, at the very end system attempted to do boot via this kexec stuff or something. that caused only some weird pixel artifacts on my system and only displaying the alt+f1 console screen or so of the booted/running 11.4 system anyways, so numlock diode was still reactive, and after several ctrl+alt+del attempts it eventually rebooted, grub shows 12.1 beta1/3.1rc7 or such entry, shows the opensuse logo screen with the little progress bar, where you can normally esc and see the messages. but that doesnt work, numlock led again seems to work when using keyboard but nothing happens. was using only usb mice/keyboards. so i was finally attaching some ps2 keyboard, and that at least allowed me to use esc at that progress bar (the bar actually never progresses anywhere, it just stays at the first few pixels for many hours now when i tried to wait for it). so using ps2 keyboard, and pressing escape all i see are those two text lines: doing fastboot creating device nodes with udev. thats all. not going anywhere from there. tried many resets (reset button of system) over and over. ctrl+alt+del dont always seem to work with this kind of haning. Reproducible: Always Steps to Reproduce: 1. 2. 3. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c zj jia <zjjia@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |zjjia@suse.com AssignedTo|bnc-team-screening@forge.pr |tonyj@suse.com |ovo.novell.com | -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c1 Tony Jones <tonyj@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |abittner@stud.fh-heilbronn. | |de --- Comment #1 from Tony Jones <tonyj@suse.com> 2011-10-17 17:56:02 UTC --- Can you reproduce with kexec outside of install system, i.e kexec -l, kexec -e to switch to a new kernel. If so, can you please try w/ latest factory kernel. Another bug was opened similar to above but the opener quickly tested using a newer kernel and said issue was resolved. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c2 andreas bittner <abittner@stud.fh-heilbronn.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|abittner@stud.fh-heilbronn. | |de | --- Comment #2 from andreas bittner <abittner@stud.fh-heilbronn.de> 2011-10-24 18:08:59 UTC --- hey there, sorry for the delay. what do you mean about that kexec stuff. that was only a brief description of the steps i did. the end result of this bug is as described, this box never boots beyond that error message. even with poweroff, cold boot. with coldboot there isnt any kexec involved is it? i know kexec only from past opensuse versions right after installing clean opensuse versions at the very end when the system wrote bootloader and final configs to the destination system (grub etc.) it would try to kexec-boot the installed new system/opensuse-kernel instead of doing a real reset/reboot/warmboot or whatever its called properly. if kexec didnt succeed, the user would then ctrl+alt+del, or reset or repower the system and the freshly installed opensuse would then boot up via normal standard means, whatever that means. so to my understanding kexec was a means to load and run a new/updated kernel right from inside the old running system without doing all the warm/cold boot-up stuff such as firmware, bios, POST etc. so anyways, as i wrote in this bug, this installed system doesnt boot up at all. it didnt boot up right after install (after the upgrade from 11.4 to 12.1) nor after the many attempts i didt be it from powered down mode, from warm reboot, cold-reboot (reset button), with usb keyboard/mice, with addtl ps2-keyboard, with ps2-keyboard only. whatever i do all it does is hang as described, and a ps2 keyboard can escape the bootscreen away and then it shows those two lines: doing fastboot creating device nodes with udev. nothing else ever happens. so i guess it has zero to do with kexec, or am i wrong? as i see today, factory has the more recent 3.1rc9 kernel as of very recently, is there any way to update this non-booting system to the most recent factory packages (rpm) in any way? maybe this hang is fixed in the 3.1rc9 kernel package? regards. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c3 --- Comment #3 from andreas bittner <abittner@stud.fh-heilbronn.de> 2011-10-27 15:18:48 UTC --- not gooooooood :( i tried to use the recently released 12.1/rc1 (x64 iso download media, booted via dvd drive) to upgrade this never-booting 12.1/beta1/x64. so yast2 from the 12.1/rc1/x64 isomedia upgraded all the packages, tried again at the end to use kexec instead of reboot, but didnt go anywhere from that screen. then i used the reset switch on the machine, for a real reboot. the grub screen comes up with two real entries for the 12.1/rc1. trying the normal entry: once again only the frog-green suse logo screen comes up and does actually quite a lot on the disk (hearing and seeing hdd i/o activity led). but doesnt boot really anywhere beyond that. usb keyboard doesnt work for ESC, but an attached ps2 keyboard works. pressing ESC brings up again the error message from beta1, but this time there is an additional error message below that: udevd.... failed to execute etc sysconfig network scripts ifup-sysctrl /etc/sysconfig/network/scripts/ifup-syscttl lo -o hotplug _ no such file or directory.... thats all. the cursor blinks there, nothing else happens. bad release :(((((( -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c4 --- Comment #4 from andreas bittner <abittner@stud.fh-heilbronn.de> 2011-10-27 15:58:21 UTC --- i strongly call for a metabug on this old generic issue of how to deal with a new product regarding a direct upgrade from a clean previous version. i think opensuse needs to make sure that a new release of opensuse must at least be upgradable and usable in some basic way coming directly from the direct previous predecessor version. i have quite often come across crazy bugs that made it impossible to upgrade a freshly set up test system with the version prior to the new release and then using the released iso image or via zypper and the other standard means that exist (see also most annoying bugs). opensuse should make sure that a clean 11.4 to the latest (11.4)+1 should work as flawlessly as possible via iso/zypper upgrade methods. if this cannot be ensured, the release should be pushed further to the future and this fundamental feature=upgrade should me enforced. its really very bothering and annoying that at so many times a very simple system using a standard clean install cannot be normally upgraded to the successor version. there are just too many bugs that have crossed my way regarding this matter in the past, and this bug seems very much like yet another bug of this kind. regards. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c5 --- Comment #5 from andreas bittner <abittner@stud.fh-heilbronn.de> 2011-10-27 16:06:25 UTC --- did a few other tests with safe-mode in the grub menu, the second entry on this 12.1/rc1/x64 system. the safemode with all those many kernel bootup parameters does boot up this 12.1/rc1/x64 system. this way i can maybe provide the previous logs, or i am going to simply attach the whole var/log/ directory or some of them at least. one thing i have figured with the safe-mode: the shutdown is very error-prone, and takes very long, at some time during the shutdown cycle after the kdm and gfx stuff is already gone and 12.1 displays those services being shutdown and so on, the whole thing comes to a stall and not much happens. at that point in time even with the ps2-keyboard and ctrl+alt+del only produces three error lines at a time right on the console alt+f1 screen or so, telling me about /dev/disk/by-id..... or dev-disk-guid..... and partitions or so errors, and eventually after quite a number of minutes the system manages to continue with the shutdown/reboot and print out some more services and status lines and even manages to reboot then. something is wrong even in this safe-mode, but i wonder where i should start hunting for the bug, most likely its some race condition of some of the devices and kernel parts or so, these shutdown problems and ctrl+alt+del might hint so. can anyone repro a clean 11.4/x64 and then directly upgrading it to 12.1/beta1 or 12.1/rc1? often i wonder why i am the one coming across these types of bugs and behaviors, one can get paranoid or really annoyed that doing supposedly easy and simple basic tasks such as upgrading a fresh current version to the latest to-be-released-version or even when it has been released, comes up with these kind of massive and barely circumventable kind of bugs :((((( regards. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c6 --- Comment #6 from andreas bittner <abittner@stud.fh-heilbronn.de> 2011-10-27 17:02:50 UTC --- wow, this situation is becoming worse the more i try to find out whats wrong. so i did a completely clean install (all partitions delete/recreate) directly with 12.1/rc1/x64 default install. the very same effect at the end of the install process.... it says its going to try kexec instead of doing real boot. after that box of text with the ****************** around it, there is actually another line telling that nscd or so is starting or shutting down or so. the disk i/o activity is actually hinting that kexec (and also during normal boots, non-failsafe normal modes) is actually trying or doing something, but nothing normal ever happens. so i waited for 10minutes at least, then tried some ctrl+alt+del on the ps2 and usb keyboards many times, waited more, nothing happens. tried to use the atx-powerbutton, sometimes the disk i/o led and activity is doing something, flickering, clicking on the disks. nothing happens. so reset button as last resort. then i try to go for failsafe-mode, thats the only mode that actually does something. as this attempt was a completely clean install of 12.1/rc1/x64, the system boots somewhat, and then suddenly crashes/exits into textmode and the textmode yast2 installer comes up and says the last attempt of install failed or so and it tries to continue this way without graphics and in some sort of fallback mode. it takes really long until it manages to activate a standard pcie/onboard realtek gigabit network device, then it does some suseconfig or whats that script called parts, with fonts and settings and so on, and then surprisingly manages to start right there back into graphics mode (still being in failsafe-boot) and presents a kde/kdm logon screen but no username/icon(?) is given, so i manually enter my testuser and password, and it brings me here, where firefox 7.0.1 works, but the green opensuse welcome screen icon doesnt do anything, but errors and complains about missing stuff on disk. so anyways, even a clean install of this 12.1/rc1/x64 is rather unusable, and this result cannot be called a normal install, and the previous attempts did not qualify for the word "upgrade". i can attach these 12.1/rc1/x64 cleaninstall logs from this system if anyone cares, and/or do some more clean-11.4/x64-to-12.1/rc1/x64 upgrade this time, maybe that gives yet again different results, but i highly doubt it. regards. p.s. things that seem odd in contrast to previous opensuse versions is, that some of the logs and during failsafe-boot (of the beta1-to-rc1 tries) and half-finished logs hint that all kinds of drivers or kernel components actually try to upload or use various nonexisting or failing firmware parts, even my realtek pcie nic has a line where some kernel part tries to upload new firmware to it, and also the ati pcie radeon hd 6970 graphics card has lines about nonexisting cayman firmware files and whatever. this sound really weird....... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c7 --- Comment #7 from Tony Jones <tonyj@suse.com> 2011-11-29 01:38:14 UTC --- Ok. We need to start again here and proceed a little more slowly and methodically. Based on comment #6 it would appear that the issue has nothing to do with updates but rather is a problem transitioning from the install kernel to the regular kernel using kexec. This is the only case of this problem reported. I realize this is frustrating but some of your previous comments are not very helpful. Please try a full-install again using the released version of 12.1. Please attach a serial console and boot the install kernel using console=ttyS{0/1}.{speed}. Please attach the console output to the bug. It was also be interesting to see if you can boot the liveCD and try kexec directly. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c Tony Jones <tonyj@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=723073 https://bugzilla.novell.com/show_bug.cgi?id=723073#c8 Tony Jones <tonyj@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |CLOSED Resolution| |NORESPONSE --- Comment #8 from Tony Jones <tonyj@suse.com> 2013-05-08 20:30:50 UTC --- No response since 2011/11/29. Product out of maintenance. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply@novell.com