Description of problem: We have a test that tries to test kdump when the system crashes. It fails to generate a vmcore and I see this in the kexec-dmesg.log. ``` Jan 27 04:27:23 localhost systemd[1]: Starting kdump-capture.service - Kdump Vmcore Save Service... Jan 27 04:27:23 localhost kdump[489]: Kdump is using the default log level(3). Jan 27 04:27:23 localhost kdump[525]: saving to /kdumproot/sysroot/ostree/deploy/fedora-coreos/var/crash/127.0.0.1-2022-01-27-04:27:23/ Jan 27 04:27:23 localhost kdump[530]: saving vmcore-dmesg.txt to /kdumproot/sysroot/ostree/deploy/fedora-coreos/var/crash/127.0.0.1-2022-01-27-04:27:23/ Jan 27 04:27:23 localhost kdump.sh[531]: Missing the log_buf symbol Jan 27 04:27:23 localhost kdump[534]: saving vmcore-dmesg.txt failed Jan 27 04:27:23 localhost kdump[536]: saving vmcore Jan 27 04:27:23 localhost kdump.sh[537]: /proc/vmcore doesn't contain vmcoreinfo. Jan 27 04:27:23 localhost kdump.sh[537]: Specify '-x' option or '-i' option. Jan 27 04:27:23 localhost kdump.sh[537]: Commandline parameter is invalid. Jan 27 04:27:23 localhost kdump.sh[537]: Try `makedumpfile --help' for more information. Jan 27 04:27:23 localhost kdump.sh[537]: makedumpfile Failed. Jan 27 04:27:23 localhost kdump[539]: saving vmcore failed, exitcode:1 Jan 27 04:27:24 localhost kdump[541]: saving the /run/initramfs/kexec-dmesg.log to /kdumproot/sysroot/ostree/deploy/fedora-coreos/var/crash/127.0.0.1-2022-01-27-04:27:23// ``` Any ideas what we're doing wrong? Version-Release number of selected component (if applicable): [root@cosa-devsh ~]# rpm -q kernel kexec-tools kernel-5.17.0-0.rc0.20220112gitdaadb3bd0e8d.63.fc36.aarch64 kexec-tools-2.0.23-4.fc36.aarch64 How reproducible: Always Steps to Reproduce: 1. rpm-ostree kargs --append='crashkernel=256M' 2. systemctl enable kdump.service 3. reboot 4. wait for kdump.service to fully come up 5. echo 1 > /proc/sys/kernel/sysrq 6. echo c > /proc/sysrq-trigger Actual results: After reboot there is no vmcore in /var/crash/ Expected results: A vmcore in /var/crash Additional info: We run a kdump test in our CI [1]. It's been failing on aarch64 in f35 [2]. I decided to try to brand new kexec-tools on rawhide and I'm seeing this new error. What we really want to do is to get the test passing in f35 and rawhide. [1] https://github.com/coreos/fedora-coreos-config/blob/testing-devel/tests/kola/kdump/crash [2] https://github.com/coreos/fedora-coreos-tracker/issues/1075#issuecomment-1022809913
This bug appears to have been reported against 'rawhide' during the Fedora 36 development cycle. Changing version to 36.
Hi Coiby, Pingfan, has this been solved? I remember this is an old issue
(In reply to Dave Young from comment #2) > Hi Coiby, Pingfan, has this been solved? I remember this is an old issue I will have a shot after a dependent bug https://bugzilla.redhat.com/show_bug.cgi?id=2052949
I have a try on the latest F36, kdump can work. [root@ampere-mtsnow-altra-10 ~]# uname -a Linux ampere-mtsnow-altra-10.khw4.lab.eng.bos.redhat.com 5.17.2-300.fc36.aarch64 #1 SMP Fri Apr 8 14:17:15 UTC 2022 aarch64 aarch64 aarch64 GNU/Linux [root@ampere-mtsnow-altra-10 ~]# rpm -qa | grep kexec-tools kexec-tools-2.0.23-6.fc36.aarch64
Per comment#4, close it. Feel free to re-open if it is not OK.