[Bug 889558] New: Kernel locks up during boot as kvm guest
https://bugzilla.novell.com/show_bug.cgi?id=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c0 Summary: Kernel locks up during boot as kvm guest Classification: openSUSE Product: openSUSE 13.1 Version: Final Platform: Other OS/Version: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: adrian@suse.com QAContact: qa-bugs@suse.de CC: mls@suse.com, trenn@suse.com, agraf@suse.com Found By: --- Blocker: --- Using our OBS appliances I do experience a lockup problem of kvm guests on a test system. The kernel seems to block while loading the initrd, but the same kernel and initrd is used for the host where it works fine. The appliance is build on top of openSUSE 13.1, including maintenance updates. The last line I see is "Unpacking initramfs", the "Freeing" line does not appear in the guests. Also, the kernel is not answering on any sysrequests at that point of time. A variation of these images are used for our official OBS workers, where such problems do not exist. So it might be a hardware specific problem. We are using a Dell system with "Intel(R) Core(TM) i5-2400 CPU". Can you give some advice to debug this further? -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c2 --- Comment #2 from Michael Schröder <mls@suse.com> 2014-07-30 11:29:41 UTC --- It locks up because kvm is started with '-m 256', but the initrd doesn't fit into this small amount of memory. A kernel panic would be preferable in this case, though... -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c3 --- Comment #3 from Adrian Schröter <adrian@suse.com> 2014-07-30 11:36:56 UTC --- This is indeed the reason. So our default value in obsworker init script is not valid, fixing this. I leave this report for a better handling in the kernel in that situation :) -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c4 --- Comment #4 from Michael Schröder <mls@suse.com> 2014-07-30 11:39:34 UTC --- The initrd also seems to be really excessive. For example, it contains python with lots of modules, but I don't see anything using it. -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c5 Adrian Schröter <adrian@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ms@suse.com --- Comment #5 from Adrian Schröter <adrian@suse.com> 2014-07-30 12:00:48 UTC --- that seems to be kiwi parts. Atm I am still not able to boot since it detects n filesystem on /dev/vda . -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c6 Thomas Renninger <trenn@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |adrian@suse.com --- Comment #6 from Thomas Renninger <trenn@suse.com> 2014-07-30 12:34:59 UTC --- Ok, this is 13.1 and does not use dracut, but still mkinitrd. In /boot there seem to be 2 initrd now: -rw-r--r-- 1 root root 85817435 Jul 30 12:16 initrd.vmx -rw------- 1 root root 7746495 Jul 30 12:05 initrd.vmx.weg The size of the 2nd looks sane. And python inside the initrd seem to be the culprit, right. This is not normal, what are you doing special there? When I do: mkinitrd -k /boot/vmlinuz -i /boot/initrd.test I get a sane in size initrd: -rw------- 1 root root 7746267 Jul 30 12:20 initrd.test -rw-r--r-- 1 root root 85817435 Jul 30 12:16 initrd.vmx -rw------- 1 root root 7746495 Jul 30 12:05 initrd.vmx.weg So this seem even not to be related to mkinitrd, but to stuff you push in extra? Adrian said something like the initrd is opened, additional stuff is added, etc... -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c7 --- Comment #7 from Michael Schröder <mls@suse.com> 2014-07-30 12:38:42 UTC --- (That initrd was generated by kiwi.) -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c8 Adrian Schröter <adrian@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|adrian@suse.com | --- Comment #8 from Adrian Schröter <adrian@suse.com> 2014-07-30 12:41:29 UTC --- We all mkinitrd, but it has actually no effect, because kiwi created the vmlinuz.vmx redirection and moved away the kernel binary. So we keep using the kiwi initrd now. I am in discussion with Marcus how to solve the issue. All what is left here for kernel people might be to handle too large initrd image nicer, not just hanging forever, but put some oops error message instead. If you don't care about this, just feel free to close the bug nevertheless :) -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c9 --- Comment #9 from Adrian Schröter <adrian@suse.com> 2014-07-30 12:41:55 UTC --- "We all" => "We call" ... -- 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=889558 https://bugzilla.novell.com/show_bug.cgi?id=889558#c10 Thomas Renninger <trenn@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX --- Comment #10 from Thomas Renninger <trenn@suse.com> 2014-07-30 12:48:52 UTC --- Hm, in fact I had oom in mind, but the screen log buffer was not big enough to scroll up to memory init logs... There is quite some more important stuff right now -> closing won't fix. -- 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