Hello everyone, I understand that this is Thanksgiving weekend so not many american eyes will see this but I still want to bounce the idea somewhere before uploading this to Debian. I recently realized that the kernel dump mechanism (kdump-tools) on debian was broken, partly because of systemd and also because of a change of format of the dmesg log. The later is fixed upstream but systemd is another issue. Debian uses a sysVinit script called kdump-tools that setup kexec and prepare things in case of a kernel panic. To make sure that a systemd enabled system does not reach multi-user when a kernel panic triggers kexec to load the new kernel, I am adding the following boot parameter : systemd.unit=kdump-tools.service The systemd kdump-tools.service unit calls the kdump-tools script that copies the content of /proc/vmcore to local storage and then reboots the server. The rest of the execution remain identical. Systems that will not be running systemd will remain unchanged. I would be happy to take comments and suggestion but keep in mind that I have limited time until debian's final freeze which is dec 5th so I will proceed with my debian upload before that. Kind regards, ...Louis -- Louis Bouchard Software engineer, Cloud & Sustaining eng. Canonical Ltd Ubuntu developer Debian Maintainer GPG : 429D 7A3B DD05 B6F8 AF63 B9C4 8B3D 867C 823E 7A61
