https://bugzilla.kernel.org/show_bug.cgi?id=219362 Michał Pecio (michal.pecio@xxxxxxxxx) changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |michal.pecio@xxxxxxxxx --- Comment #17 from Michał Pecio (michal.pecio@xxxxxxxxx) --- These operations are asynchronous and they don't even seem to be synchronized against each other. This is what I got from s followed u: [ 9oct 19:21] sysrq: Emergency Sync [ +0,000028] sysrq: Emergency Remount R/O [ +0,019142] EXT4-fs (sda2): re-mounted ef717e3d-539d-4988-9a02-ca541d643379 ro. Quota mode: disabled. [ +0,002201] Emergency Sync complete --- some long error message here Emergency Remount complete If subsequent o would execute before the u is done, the u would only increase risk of data loss by making the disk busy right before powerdown and by exercising the remount code against a concurrent loss of power, which might be a neglected case due to its rarity. You will never know whom to blame without hours of debugging. Myself, I would try sync ; poweroff -f -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.