Hi,
I have two questions about entering emergency mode when mount target fails. If local-fs.target fails in booting stage, it will come into emergency mode and wait for user interaction. In this failure case, is it better to introduce a crash generate core dump and reboot automatically? Replacing emergency mode by such a method can be useful when local-fs.target's failure is transient or user interaction is not convenient. A possible crash service may like this: [Unit] DefaultDependencies=no Description=Crash the system using the SysRq trigger [Service] ExecStartPre=-/bin/sh -c 'echo "crash-system.service started" > /dev/kmsg' ExecStart=/bin/sh -c 'echo c > /proc/sysrq-trigger' Type=oneshot And this new service(instead of current emergency.target) is used in OnFailure of local-fs.target. Does this sound good and useful? Or what is the drawback of this solution? Another question is: is there a configuration option to cause a reboot if local-fs.target fails? So that we can use the option directly. Thank you in advance. Best regards, Yang
_______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/systemd-devel
