[Bug 1207114] New: kdump doesn't work (at least in 6.1) -- dracut time outs
https://bugzilla.suse.com/show_bug.cgi?id=1207114 Bug ID: 1207114 Summary: kdump doesn't work (at least in 6.1) -- dracut time outs Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel Assignee: jbohac@suse.com Reporter: jslaby@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- Created attachment 864075 --> https://bugzilla.suse.com/attachment.cgi?id=864075&action=edit /run/initramfs/rdsosreport.txt dracut from crash-kernel ends with this:
Warning: dracut-initqueue: timeout, still waiting for following initqueue hooks: Warning: /lib/dracut/hooks/initqueue/finished/devexists-\x2fdev\x2fdisk\x2fby-uuid\x2fcf48897f-62c1-4314-ad2f-75e9ff0b32ed.sh: "[ -e "/dev/disk/by-uuid/cf48897f-62c1-4314-ad2f-75e9ff0b32ed" ]" Warning: dracut-initqueue: starting timeout scripts ... Warning: dracut-initqueue: starting timeout scripts Warning: Could not boot.
I am attaching a full log. I don't know if it is my setup or something else. But I could not any crash output even on my bare metal. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207114 https://bugzilla.suse.com/show_bug.cgi?id=1207114#c1 --- Comment #1 from Jiri Slaby <jslaby@suse.com> --- Created attachment 864076 --> https://bugzilla.suse.com/attachment.cgi?id=864076&action=edit lsinitrd /boot/initrd-kdump Worth noting (in the dracut emergency shell): # ls -l /dev/disk ls: cannot access '/dev/disk': No such file or directory # ls -l /dev/[sv]d* ls: cannot access '/dev/[sv]d*': No such file or directory So there are no disks whatsoever. Nor there are any loaded modules. Ah, but the loaded /boot/initrd-kdump contains 6.1.3-rc1 modules (see the attachment), but the crash kernel booted is actually 6.1.4-rc1. This looks like /boot/initrd-kdump was not properly regenerated during kdump.service. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207114 https://bugzilla.suse.com/show_bug.cgi?id=1207114#c2 --- Comment #2 from Jiri Slaby <jslaby@suse.com> --- rm /boot/initrd-kdump fixed this. So something didn't update the initrd properly. I still have the VM in the original (buggy) state if more data are needed. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207114 https://bugzilla.suse.com/show_bug.cgi?id=1207114#c3 --- Comment #3 from Jiri Slaby <jslaby@suse.com> --- In: sh -xx /usr/lib/kdump/load.sh --update I see: ++ mkdumprd -K /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default -I /boot/initrd-kdump + output='>>> Not regenerating kdump initrd. Use mkdumprd -f to force regeneration.' So: sh -xx mkdumprd -K /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default -I /boot/initrd-kdump ends with: + '[' -z /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default ']' + '[' -f /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default ']' + '[' -z /boot/initrd-kdump ']' + (( ! 0 )) + '[' -f /boot/initrd-kdump ']' + '[' /etc/sysconfig/kdump -ot /boot/initrd-kdump ']' + '[' /root/.ssh/id_dsa.pub -ot /boot/initrd-kdump ']' + '[' /root/.ssh/id_dsa -ot /boot/initrd-kdump ']' + '[' /etc/hosts -ot /boot/initrd-kdump ']' + '[' /etc/nsswitch.conf -ot /boot/initrd-kdump ']' + '[' /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default -ot /boot/initrd-kdump ']' + status_message 'Not regenerating kdump initrd. Use mkdumprd -f to force regeneration.' + local 'msg=Not regenerating kdump initrd. Use mkdumprd -f to force regeneration.' + (( ! 0 )) + echo '>>> Not regenerating kdump initrd. Use mkdumprd -f to force regeneration.' >>> Not regenerating kdump initrd. Use mkdumprd -f to force regeneration. + exit 0 So: # ls -l /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default /boot/initrd-kdump /boot/../usr/lib/modules/6.1.4-rc1-1.gbc1c341-default/vmlinuz -rw-r--r-- 1 root root 11079232 Jan 5 08:48 /boot/../usr/lib/modules/6.1.4-rc1-1.gbc1c341-default/vmlinuz -rw------- 1 root root 19468756 Jan 5 12:22 /boot/initrd-kdump lrwxrwxrwx 1 root root 55 Jan 5 12:23 /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default -> ../usr/lib/modules/6.1.4-rc1-1.gbc1c341-default/vmlinuz Ah, you are testing a link age. This is apparently fallout of move from /boot/ to /usr/lib. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207114 https://bugzilla.suse.com/show_bug.cgi?id=1207114#c4 --- Comment #4 from Jiri Slaby <jslaby@suse.com> --- (In reply to Jiri Slaby from comment #3)
+ '[' /boot/vmlinuz-6.1.4-rc1-1.gbc1c341-default -ot /boot/initrd-kdump ']'
The check doesn't make sense at all for packaged kernels, IMO. vmlinuz in /usr (the link target) have some arbitrary time which is stored in rpm. The link above have some random time, set by installation time. -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠suse.com