Fix "E: dpkg was interrupted, you must manually run 'sudo dpkg --configure -a'"

sudo dpkg --configure -a
sudo apt -f install
sudo apt update && sudo apt upgrade

What the error means

Any apt command (install, upgrade, even apt autoremove) refuses to run and prints:

E: dpkg was interrupted, you must manually run 'sudo dpkg --configure -a' to correct the problem.

Installing a .deb happens in two stages: unpack (copy files into place) and configure (run the package's post-install script, set up services, rebuild caches). dpkg records each package's state in /var/lib/dpkg/status and keeps a journal in /var/lib/dpkg/updates/. If the process stops between or during those stages, packages are left "unpacked" or "half-configured" and apt will not start a new transaction until the old one is finished. dpkg --configure -a resumes it.

Cause 1: the upgrade was interrupted

The usual story: the laptop ran out of battery, you closed the terminal or SSH session, you pressed Ctrl+C, or you rebooted while "Software Updater" or unattended-upgrades was running. In all these cases the quick fix above is the complete solution. Run it from a terminal you will not close; over SSH, start tmux or screen first so a dropped connection does not interrupt it again.

Check that everything is consistent afterwards:

sudo dpkg --audit          # lists packages still not fully installed
dpkg -l | grep -v '^ii'    # anything not in the normal "ii" state

No output from dpkg --audit means the database is clean. Lines starting with iU (unpacked) or iF (half-configured) are still broken.

Cause 2: a package fails to configure

If dpkg --configure -a ends with something like:

dpkg: error processing package nvidia-dkms-580 (--configure):
 installed nvidia-dkms-580 package post-installation script subprocess returned error exit status 10
Errors were encountered while processing:
 nvidia-dkms-580

then the interruption was not the whole story: one package's setup script fails. Scroll up to the first error line for that package; it usually says why. The most common cases are a DKMS kernel module that does not build against the new kernel, a service that fails to start, and a full disk. These are covered in detail on the "dpkg returned an error code (1)" page. Fix the underlying cause, then run sudo dpkg --configure -a again.

Cause 3: the disk or /boot is full

Kernel updates write a new kernel and initramfs to /boot. If /boot is a small separate partition, or the root filesystem is full, the update dies part-way through. Check:

df -h / /boot /boot/efi

If /boot is at or near 100%, remove old kernels you are not booted into (check yours with uname -r), then resume:

dpkg -l 'linux-image-*' | grep ^ii
sudo apt purge linux-image-6.8.0-31-generic   # an old one, never the running kernel
sudo dpkg --configure -a

If apt refuses to purge because of the interrupted state, use sudo dpkg --purge on the old kernel image packages directly. For a full root filesystem, see "No space left on device".

Cause 4: dpkg is waiting for input

If dpkg --configure -a appears to hang, it is often waiting for you. A configuration file prompt looks like this, and may be hidden behind other output:

Configuration file '/etc/ssh/sshd_config'
 ==> Modified (by you or by a script) since installation.
 ==> Package distributor has shipped an updated version.
   What would you like to do about it ?  Your options are:
    Y or I  : install the package maintainer's version
    N or O  : keep your currently-installed version
 The default action is to keep your current version.
*** sshd_config (Y/I/N/O/D/Z) [default=N] ?

Answer it (pressing Enter keeps your version). Kernel module builds through DKMS can also take several minutes with no output. To see what is actually running:

ps -ef --forest | grep -A6 '[d]pkg'

Cause 5: the system will not boot after the interruption

If the interruption happened during a kernel, systemd or libc update, the desktop may not start. Boot an older kernel or recovery mode from the GRUB menu (hold Shift on BIOS machines or press Esc on UEFI during boot), open a root shell and run:

mount -o remount,rw /
dpkg --configure -a
apt -f install
update-grub
reboot

Ubuntu 26.04 builds the initramfs with Dracut. If the kernel package failed while generating it, regenerate it after the repair with sudo dracut --force --kver $(uname -r), or for every installed kernel with sudo dracut --force --regenerate-all. Debian 13 and Ubuntu 24.04 use initramfs-tools: sudo update-initramfs -u -k all. If GRUB itself is damaged, see the GRUB rescue guide.

Distro-specific notes

• Ubuntu release upgrades: if do-release-upgrade was interrupted, run the quick fix, then sudo apt full-upgrade to complete the remaining packages. The Ubuntu upgrade guide lists post-upgrade checks.
• Linux Mint: the Update Manager shows the same error in a dialog; the terminal commands are identical.
• Fedora (DNF5) and openSUSE use RPM, which does not have a separate configure stage. After an interrupted transaction, run sudo dnf distro-sync or sudo zypper verify. On Arch, rerun sudo pacman -Syu; if it complains about a lock, see pacman "unable to lock database".

How to prevent it

• Plug in laptops before large upgrades and do not close the updater window.
• Run upgrades over SSH inside tmux or screen.
• Keep /boot tidy with sudo apt autoremove, which removes old kernels automatically.
• Do not kill apt or dpkg to get rid of a lock; see "Could not get lock" for the safe way.

Frequently asked questions