Confirmed fix — two root causes identified and resolved on two ZFS-root machines
I can confirm this bug and have identified the root causes. I have two machines (a Kingdel mini PC and a Lenovo Yoga 9i laptop) both running Ubuntu 25.10 with encrypted ZFS root. Both were stuck at sockets.target on any kernel newer than 6.17.0-6-generic (tested up to 6.17.0-20-generic). Both now boot 6.17.0-20-generic cleanly after the fix below. Environment (identical on both machines): • Ubuntu 25.10 (Questing Quokka) • Root filesystem: ZFS (encrypted rpool) • OpenZFS: 2.3.4-1ubuntu2 • systemd: 257.9-0ubuntu2.3 • kdump-tools: 1:1.10.7ubuntu2 (USE_KDUMP=1) • 8× CIFS automounts in /etc/fstab using x-systemd.automount Root Cause #1 — kdump-tools (primary): USE_KDUMP=1 causes kdump to prepare a crash kernel early in boot. On 6.17.0-7+, this interacts badly with encrypted ZFS root, deadlocking the systemd boot sequence. This aligns with Bug #2129972 and the AskUbuntu thread at https://askubuntu.com/questions/1563612. Root Cause #2 — CIFS automount ordering (contributing): CIFS entries in /etc/fstab using x-systemd.automount without x-systemd.mount-timeout=10s can cause systemd PID 1 to enter D state (uninterruptible sleep) if the network isn't ready when the automount unit triggers during early boot. This matches Lee Walsh's original finding (comment #4) and is also documented in Bug #2142726. Fix applied (both machines now boot 6.17.0-20-generic): 1. Disable kdump: sudo sed -i 's/^USE_KDUMP=1/USE_KDUMP=0/' /etc/default/kdump-tools 2. Add mount timeouts to all CIFS entries in /etc/fstab: Ensure every CIFS line has x-systemd.idle-timeout=60,x-systemd.mount-timeout=10s,nofail,_netdev. If your entries already have x-systemd.automount but are missing timeouts, this sed one-liner adds them: sudo sed -i '/cifs/{/x-systemd.mount-timeout/! s/x-systemd.automount,/x-systemd.automount,x-systemd.idle- timeout=60,x-systemd.mount-timeout=10s,/}' /etc/fstab 3. Update GRUB to boot the latest kernel: sudo sed -i 's/^GRUB_DEFAULT=.*/GRUB_DEFAULT=0/' /etc/default/grub sudo update-grub 4. Reboot — should boot into 6.17.0-20 (or latest) cleanly. To revert if anything goes wrong, restore backups and boot the old 6.17.0-6 kernel from GRUB. Related bugs: • Bug #2129972 — kdump-tools prevents boot on encrypted ZFS • Bug #2142726 — systemd PID 1 hangs on CIFS fstat • Bug #2130843 — automount blocking with rust-coreutils stat Summary: If you have encrypted ZFS root + kdump enabled + CIFS automounts in fstab, you likely need both fixes. Disabling kdump alone may be sufficient for some, but adding CIFS mount timeouts provides defence in depth against the systemd ordering race. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2132966 Title: new kernel 6.17.7 fails to boot To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2132966/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
