All my systems have an HDD drive. I am going to try it on an SSD system. I am unable to figure out why bookworm has no issue but trixie does performing the exact same steps.
On Thu, Nov 7, 2024 at 4:37 PM Theodore Ts'o <ty...@mit.edu> wrote: > On Thu, Nov 07, 2024 at 11:08:19AM -0500, Anees Ahmad wrote: > > Hello Ted, > > > > I have tried on 4 different computers and it is reproducible on all of > them > > on unclean shutdown. > > Well, I can't reproduce it for on my Debian Trixie. I'm running on a > Dell Precision 3640 Tower, with an NVMe SSD. Kernel version is > linux-image-6.11.5-amd64. E2fsprogs version was 1.47.1-1+b1. > > My experiment was as follows. I did a "make clean; make" in a Linux > kernel tree. So the ext4 file system was getting actively modiied. > Then in a root shell I ran "echo b > /proc/syrq-trigger" which trigers > an immediate reboot without syncing or umounting the file systems. > > During the reboot, the journal got replayed, including processing > about two dozen orphaned inodes. With no problem. > > So, "it works for me", and I can work a problem that is not > reproducible. And I haven't heard anyone else complain, either on > other Linux distributions, or any other Debian users. If this was so > easily reproucible, why isn't any one else seeing this? > > Is there anything else "interesting" about your systems? > > - Ted >
