apt 404 Not Found
apt asked for a file the server no longer has. Usually stale package lists: run apt update and retry.
Find the exact message you're seeing. Each page explains the cause in plain words and gives the fix first.
Tip: use the search box at the top of the page and paste the part of the message that doesn't change between machines — for example Unable to locate package rather than the package name.
apt asked for a file the server no longer has. Usually stale package lists: run apt update and retry.
Another apt/dpkg process holds the lock: find it, let it finish, then repair.
dnf can't fetch a repo's index: fix network/DNS, clean the cache, or disable the broken repo.
A package's install script failed. Find which one and why, then repair it.
An install was cut off: run dpkg --configure -a, then fix whatever still fails.
apt can't find the package: refresh lists, check the name, enable universe/contrib.
A repository's signing key is missing: install it as a keyring with signed-by, not apt-key.
Outdated keyring or bad cache: update archlinux-keyring, then repopulate with pacman-key.
A stale or active pacman lock: make sure nothing runs, remove db.lck, rerun the upgrade.
pip refuses to touch system Python: use a venv, pipx, or the distro package.
apt can't find the repo's index: wrong release, unsupported PPA, or an end-of-life system.
The snap command cannot reach the snapd daemon. Start or install snapd, or check WSL, containers and Mint.
apt can't satisfy dependencies: find the real blocker, then fix PPAs, holds or mixed releases.
Disk full, inodes exhausted or deleted-but-open files: find the cause and free space safely.
Disk suddenly read-only? Read dmesg, repair with fsck, check the drive and remount read-write.
File system metadata is corrupt: unmount, check hardware, repair with xfs_repair or e2fsck.
Find which process holds the mount with lsof or fuser, stop it, then unmount cleanly.
Missing driver or a container layer (LVM, LUKS, RAID) that must be opened before mounting.
Usually the wrong device, a missing CIFS/NFS helper or a bad option, not corruption.
RAM and swap exhausted, strict overcommit, or a map/address-space limit. Diagnose and fix.
Usually a missing or broken initramfs after a kernel update. Boot an older kernel and rebuild it.
Process killed by the OOM killer? Confirm it in the logs, find the memory hog, add swap and set limits.
Find why a program segfaults with coredumpctl and gdb, then fix packages, configs, RAM or your code.
Booted into emergency mode? Read the journal, fix the bad fstab line or run fsck, then reboot.
ssh could not confirm the server's identity and had no way to ask you. Add the verified key to known_hosts.
Your own machine has no route for the address. Usually no default gateway, a down link, or IPv6 without IPv6.
Either the host does not answer on the local network or a firewall rejects you. Tell them apart in two steps.
SSH rejected every key you offered. Find out which key, which user and which file is wrong.
The server's host key no longer matches known_hosts. Verify it, then remove the stale entry.
Nothing accepted the connection on that port. Install or start sshd, then check port and firewall.
Your SYN got no answer at all. Find where packets are dropped: wrong IP, host down, firewall or NAT.
Your machine's own name is not in /etc/hosts. Add one line and the warning and the delay go away.
Your system can't reach a DNS server. Check the network, resolv.conf and systemd-resolved.
The script lacks the execute bit or lives on a noexec mount. chmod +x fixes most cases.
Your account has no sudo rights. Add it to the sudo or wheel group from a root shell.
Windows CRLF line endings broke the shebang: strip them with sed or dos2unix.
A glob expanded to more arguments than the kernel accepts. Use find or xargs instead.
The shell can't find the program. Install the right package or add its directory to PATH.
A program needs a .so library the loader can't find: install it or register its path.
Wrong CPU architecture, a non-Linux binary or a script without a shebang.
The program was built against a newer glibc than yours. Never replace system glibc.
EPERM even as root: immutable attribute, mount options, sticky bit or missing capabilities.
A process hit its file descriptor limit. Raise the right limit, then check for a leak.
Find which process holds the port with ss or lsof, then stop it, change ports or fix the restart.
systemctl cannot reach systemd: no systemd as PID 1 (WSL, container, chroot) or no user bus after su.
A systemd service won't start? Read its journal, find the real error and fix config, port or permissions.
systemd has no unit by that name. Usually a different name on your distro, a missing package or no daemon-reload.
Your user can't access docker.sock. Join the docker group (and re-login) or use rootless mode.
The repo is owned by another user. Fix ownership, or add that one path to safe.directory.
The NVIDIA kernel module is not loaded. Usually a kernel update without a rebuilt module, or Secure Boot.