How to fix Permission denied
chmod / chown / sudo usage and avoiding common pitfalls.
View solution →🩺 Troubleshooting
A reverse-lookup reference to quickly diagnose Linux errors and incidents by symptom. Covers Permission denied, No space left on device, SSH failures, high CPU load, memory pressure, DNS resolution failures, and more.
🔒 Permission denied
PermissionsFixing "Permission denied (publickey)"
Check whether the key is offered, authorized_keys, permissions, and sshd_config from both client and server sides.
View solution →🌐 Connectivity / Network
NetworkSSH connection troubleshooting checklist
Permission denied (publickey), Host key verification failed, Connection timeout.
View solution →Fixing SSH Disconnects: Keepalive and Timeouts
When SSH drops after a pause, triage idle timeouts vs missing keepalive and fix it for good with ServerAlive / ClientAlive.
View solution →Fixing "Too many authentication failures"
Understand how offering too many ssh-agent keys exceeds MaxAuthTries, and fix it for good with IdentitiesOnly and ~/.ssh/config.
View solution →Fixing "Host key verification failed"
When a rebuilt server or reused IP makes the known_hosts key mismatch, remove the entry with ssh-keygen -R and reconnect safely.
View solution →DNS resolution troubleshooting
dig / host / nslookup usage and /etc/resolv.conf inspection.
View solution →Port connectivity troubleshooting
nc / telnet / curl for connection check, ss / netstat for port state.
View solution →Fixing "Address already in use"
Find the process holding the port with ss / lsof / fuser, release it safely, and handle TIME_WAIT and double-start cases.
View solution →Diagnosing "Connection refused"
The host is reachable but actively rejects you. Tell a stopped service from a blocked port, and refused from timeout, in the fewest steps.
View solution →Diagnosing "No route to host"
There is no path to deliver packets to the host. Triage a missing route, failed ARP, a router, and a firewall REJECT, from your host outward.
View solution →Diagnosing "Connection timed out"
A silent, long wait with no reply. Triage a firewall DROP, an unopened security group, a dead service, and a path blackhole, from your host outward.
View solution →Fixing "certificate verify failed"
Triage the SSL verification error in curl and Python with openssl s_client, then fix CA bundle, incomplete chain, or clock skew.
View solution →Fixing "sudo: unable to resolve host"
Treat the unable to resolve host warning as a mismatch between hostname and /etc/hosts, then fix the 127.0.1.1 line to clear both the warning and the delay.
View solution →📊 Server Performance
PerformanceDiagnosing 100% CPU usage
Identifying culprit processes with top / ps / load average.
View solution →Diagnosing high load average
Tell CPU-bound from I/O-bound when load average is high. How to read uptime/top/vmstat/iostat to find the cause.
View solution →Handling OOM killer events
When a process is killed without explanation. Confirm with dmesg/journalctl, adjust oom_score_adj, and add swap.
View solution →Fixing "Too many open files" - Resolving File Descriptor Exhaustion
"Too many open files" means you hit the file descriptor limit. Use ulimit, lsof, limits.conf and sysctl to diagnose and fix it permanently.
View solution →Understanding and Handling Zombie Processes in Linux
Zombie processes are terminated processes that linger in the process table. Learn what causes them, how to find them, and how to clean them up safely.
View solution →Fixing "fork: Resource temporarily unavailable"
Diagnose why fork fails with EAGAIN by checking the per-user ulimit, the cgroup TasksMax, and the system pid_max in order, then fix it permanently.
View solution →Diagnosing Swap Thrashing
Spot swap thrashing with vmstat si/so when a server suddenly slows, find the culprit process, and fix it permanently with swappiness and cgroups.
View solution →Fixing "Cannot allocate memory"
Fix Cannot allocate memory (ENOMEM) even when RAM looks free by diagnosing overcommit accounting, ulimit, and max_map_count.
View solution →💾 Disk Space
DiskIdentifying Docker disk usage
docker system df to break down image / container / volume usage.
View solution →Handling inode exhaustion
"No space left on device" while df -h shows free? Diagnose with df -i and recover.
View solution →Fixing "Argument list too long"
Why rm * fails and how to fix it with find -delete and find + xargs.
View solution →Fixing "Read-only file system"
A sudden read-only mount. Triage with findmnt/dmesg, then decide fast whether to remount,rw or run fsck.
View solution →Fixing "Stale file handle" on NFS
What an NFS Stale file handle (ESTALE) really is. Triage with findmnt/lsof, fix it with a remount, and pin fsid to stop it recurring.
View solution →Disk Full but df Shows Space
A process holding a deleted file open, or inode exhaustion. Split the two-way gap with lsof +L1 / df -i.
View solution →Fixing "device is busy" on umount
When umount fails with busy, find the process holding the filesystem with fuser/lsof and detach it safely. Covers lazy umount, cwd, and swap pitfalls.
View solution →Fixing "Text file busy"
A running binary can't be overwritten with cp. Find the process with lsof and swap it in safely with rename.
View solution →Diagnosing "Input/output error"
A file op returns EIO. Read the kernel log with dmesg, then tell a failing disk from FS corruption or a disconnect.
View solution →Repairing Filesystem Corruption with fsck
Repair ext4 corruption safely. Why you must unmount first, how to check the root FS, -n/-y/-f options, and the XFS difference.
View solution →📜 Web Server / Logs
LogsReading Nginx/Apache logs
access.log / error.log paths and how to read incident causes.
View solution →Fixing Server Time Skew - Synchronizing NTP with chrony and timedatectl
How to fix server time skew on Linux. Use timedatectl status to check NTP state, chronyc tracking to diagnose drift, and chronyc makestep to force immediate sync.
View solution →Fixing Clock Skew and Certificate Errors
Triage "Clock skew detected" build warnings and "certificate is not yet valid" from a wrong clock, verify with date/openssl, and fix it for good.
View solution →🐚 Script Execution Errors
ShellFixing "bad interpreter"
A script fails with bad interpreter. Tell CRLF line endings from a wrong shebang path using file/cat -A, then fix it with dos2unix/sed.
View solution →When Cron Jobs Don't Run
Diagnose why a cron job won't run fast. A checklist that walks the daemon, logs, PATH, environment, schedule syntax, and permissions in order.
View solution →"command not found" for Your Own Scripts
Fix command not found for a script you wrote. Tell it apart from Permission denied, then run with ./, chmod +x, add to PATH, and clear the hash cache.
View solution →Decoding "syntax error near unexpected token"
Read the bash syntax error near unexpected token message correctly. Understand what the token points to, then debug fi/then/done, (, newline, and end of file.
View solution →Fixing "cannot set LC_ALL" Locale Warnings
Triage cannot set LC_ALL / setlocale warnings from a missing locale with locale -a, generate it with locale-gen, and stop SSH forwarding for good.
View solution →"No such file or directory" When the File Exists
ls shows the file, yet you get No such file or directory. Isolate hidden chars, CRLF, broken links, and a missing dynamic linker.
View solution →⚙️ Services / systemd
ServiceDebugging "Failed to start" systemd Services
Diagnose systemctl "Failed to start" fast. Work through status, journalctl, and exit codes to fix ExecStart, permissions, dependencies, and start-limit issues.
View solution →📦 Package Management / apt
PackageFixing "Could not get lock"
When apt stalls with "Could not get lock", find the process holding the lock, release it safely, and recover an interrupted dpkg - in the fewest steps.
View solution →Fixing Broken Dependencies and Held Packages
When apt holds updates back or stops with unmet dependencies, learn to tell held from broken and repair the dependency tree without making it worse.
View solution →Fixing Repository GPG Key Errors (NO_PUBKEY)
Tell a NO_PUBKEY error from an expired EXPKEYSIG one, fetch the public key into a keyring, bind it with signed-by, and restore apt signature verification safely.
View solution →🚀 Startup / Boot
SystemWhen Linux Won't Boot - Kernel Panic and Emergency Mode
First steps when Linux won't boot. Triage kernel panic, the initramfs prompt, GRUB and systemd emergency mode, then recover with fstab fixes, fsck and an older kernel.
View solution →🐧 Practice your skills hands-on
Reading is just the start — practice in the browser-based terminal to build muscle memory.