Am Montag, dem 21.04.2025 um 13:57 +0200 schrieb Friedrich Romstedt: > Hi, > > Yesterday I prepared an USB stick with the recent version of the Arch > Linux installation medium (Release 2025-04-01). > > This is about some error entries in the system journal, which I > encountered during copying this ISO image over. The respective > transcripts are provided inline, as well as attached in the form of > plain text files.
See https://github.com/systemd/systemd/issues/36624 and https://github.com/systemd/systemd/pull/36685 It appears that current udevd synchronously probes partition table of USB device. Your device seems to be too slow to answer in time while under write load, so udevd stalls in the ioctl syscall and fails to respond to the watchdog in time. As a consequence, systemd then kills the procees. It look as if this issue was fixed in March (see above), but that fix hasn't found its way into a release yet. Cheers, Basti > > I used initially $ sudo cp <ISO> /dev/sda. After having finished that > process, I spotted some error log entries in the system journal: > > Apr 20 15:30:25 T470p sudo[21475]: friedrich : TTY=pts/2 ; > PWD=/home/friedrich/(...) ; USER=root ; COMMAND=/usr/bin/cp > archlinux-2025.04.01-x86_64.iso /dev/sda > Apr 20 15:30:25 T470p sudo[21475]: pam_unix(sudo:session): session > opened for user root(uid=0) by friedrich(uid=1000) > Apr 20 15:33:10 T470p systemd[1]: systemd-udevd.service: Watchdog > timeout (limit 3min)! > Apr 20 15:33:10 T470p systemd[1]: systemd-udevd.service: Killing > process 359 (systemd-udevd) with signal SIGABRT. > Apr 20 15:34:40 T470p systemd[1]: systemd-udevd.service: State 'stop- > watchdog' timed out. Killing. > Apr 20 15:34:40 T470p systemd[1]: systemd-udevd.service: Killing > process 359 (systemd-udevd) with signal SIGKILL. > Apr 20 15:36:10 T470p systemd[1]: systemd-udevd.service: Processes > still around after SIGKILL. Ignoring. > Apr 20 15:37:38 T470p kernel: sda: sda1 sda2 > Apr 20 15:37:38 T470p sudo[21475]: pam_unix(sudo:session): session > closed for user root > Apr 20 15:37:38 T470p systemd[1]: systemd-udevd.service: Main process > exited, code=killed, status=9/KILL > Apr 20 15:37:38 T470p systemd[1]: systemd-udevd.service: Failed with > result 'watchdog'. > Apr 20 15:37:38 T470p systemd[1]: systemd-udevd.service: Consumed > 4.845s CPU time, 51.6M memory peak. > Apr 20 15:37:38 T470p systemd[1]: systemd-udevd.service: Scheduled > restart job, restart counter is at 1. > Apr 20 15:37:38 T470p systemd[1]: Starting Rule-based Manager for > Device Events and Files... > Apr 20 15:37:39 T470p systemd-udevd[21582]: Using default interface > naming scheme 'v257'. > Apr 20 15:37:39 T470p systemd[1]: Started Rule-based Manager for > Device Events and Files. > Apr 20 15:38:19 T470p kernel: usb 1-1: USB disconnect, device number > 7 > > Because I had doubts about the integrity of the thus prepared USB > stick, I repeated the process, this time using $ su - and # dd > if=<ISO> of=/dev/sda bs=8M. Here, I encountered similar logs: > > Apr 20 15:38:53 T470p kernel: sda: sda1 sda2 > Apr 20 15:38:53 T470p kernel: sd 0:0:0:0: [sda] Attached SCSI > removable disk > Apr 20 15:39:07 T470p su[21650]: (to root) friedrich on pts/2 > Apr 20 15:39:07 T470p su[21650]: pam_unix(su:session): session opened > for user root(uid=0) by friedrich(uid=1000) > Apr 20 15:44:05 T470p systemd[1]: systemd-udevd.service: Watchdog > timeout (limit 3min)! > Apr 20 15:44:05 T470p systemd[1]: systemd-udevd.service: Killing > process 21582 (systemd-udevd) with signal SIGABRT. > Apr 20 15:45:35 T470p systemd[1]: systemd-udevd.service: State 'stop- > watchdog' timed out. Killing. > Apr 20 15:45:35 T470p systemd[1]: systemd-udevd.service: Killing > process 21582 (systemd-udevd) with signal SIGKILL. > Apr 20 15:47:05 T470p systemd[1]: systemd-udevd.service: Processes > still around after SIGKILL. Ignoring. > Apr 20 15:48:35 T470p systemd[1]: systemd-udevd.service: State > 'final-sigterm' timed out. Killing. > Apr 20 15:48:35 T470p systemd[1]: systemd-udevd.service: Killing > process 21582 (systemd-udevd) with signal SIGKILL. > Apr 20 15:48:56 T470p kernel: sda: sda1 sda2 > Apr 20 15:48:56 T470p systemd[1]: systemd-udevd.service: Main process > exited, code=killed, status=9/KILL > Apr 20 15:48:56 T470p systemd[1]: systemd-udevd.service: Failed with > result 'watchdog'. > Apr 20 15:48:56 T470p systemd[1]: systemd-udevd.service: Consumed > 1.101s CPU time, 7.4M memory peak. > Apr 20 15:48:56 T470p systemd[1]: systemd-udevd.service: Scheduled > restart job, restart counter is at 2. > Apr 20 15:48:56 T470p systemd[1]: Starting Rule-based Manager for > Device Events and Files... > Apr 20 15:48:56 T470p systemd-udevd[21808]: Using default interface > naming scheme 'v257'. > Apr 20 15:48:56 T470p systemd[1]: Started Rule-based Manager for > Device Events and Files. > Apr 20 15:49:13 T470p kernel: usb 1-1: USB disconnect, device number > 8 > > In the second run, where I kept an eye on the journal, I noticed that > the error entries were logged a few minutes before the copying > process completed. There were no errors reported by both of these > directly. > > The pen drive is booting and working just fine. However, I felt this > behaviour being worth it to be reported. Maybe someone else can make > sense of it. > > Best, > Friedrich
