Start by finding the failure boundary, not by rebooting. From a separate network, test DNS and HTTP/TCP reachability. If the host itself is unreachable, switch to your provider’s serial, KVM, or cloud console and preserve logs before taking action. Once you have access, use systemctl and targeted journalctl queries to determine whether a service failed, memory and swap were exhausted, startup dependencies broke, or the kernel crashed. Persistent journald storage and kdump evidence are what let you distinguish a reboot from a hard crash and fix the underlying cause.
Need a server to try this on? Antihost offers Linux VPS plans with NVMe storage, full root access and a 14-day free trial.
1. Establish exactly what is down
A browser error alone does not prove that a Linux server crashed. A DNS record, firewall, load balancer, web server, application, database, or the host itself can fail while producing a similar symptom. Test from a network that is independent of the server’s own connection so a local routing or resolver problem does not mislead you.
As an Amazon Associate I earn from qualifying purchases.
Check DNS resolution
Query the authoritative result and note the returned address, record type, and response time:
dig +short example.com A
dig +short example.com AAAA
dig example.comIf the expected address is missing or has changed, investigate DNS before logging into the machine. A valid answer only proves that a resolver returned a record; it does not prove that the host is reachable.
#1 Best Overall
- [Ultra-Lightweight Ear Buds Designed for Small Ears] Each earbud weighs only 3.7g and features a compact, ergonomic in-ear design made especially for small ears. Secure, low-profile, and comfortable for workouts, all-day wear or overnight sleeping.The ultra-slim profile sits flush in your ear with no hard edges poking or rubbing, so you can even sleep on your side without discomfort.
- [Immersive Stereo Sound with TOZO OrigX Technology] TOZO OrigX tuning delivers clear vocals, balanced mids, and natural stereo sound for music, podcasts, and videos.
- [Long Battery Life for Daily Use] Get up to 7 hours of playtime on a single charge, with up to 32 hours total using the charging case—ideal for workdays, commuting, and extended listening sessions.
- [Bluetooth 5.3 & Stable Connection] Bluetooth 5.3 provides fast pairing, stable wireless performance, and reduced dropouts as you move around home or office.
- [Deep Bass with Clear Vocals] High-performance drivers produce punchy bass while keeping vocals clean and detailed for everyday listening.
Test TCP and HTTP separately
Probe the ports that should be exposed. A successful TCP connection with an HTTP error points to the web stack or application; a refused connection often means nothing is listening; a timeout can indicate a host, route, firewall, or provider problem.
nc -vz example.com 80
nc -vz example.com 443
curl -I --max-time 10 https://example.com
curl -sS -o /dev/null -w '%{http_code} %{time_total}n' https://example.comRepeat the checks against the server’s IP address when appropriate, while accounting for virtual-host configuration and TLS certificate matching. Compare HTTP, SSH, and any other critical service. If HTTP fails but SSH works, begin with the web and application layers. If SSH and HTTP both fail, treat the incident as a host, network, or infrastructure failure until the console proves otherwise.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Record the first observed time
Write down when independent checks first failed, the source network, resolved addresses, status codes, and whether failures were timeouts or refusals. This timestamp becomes the window for journal and provider-event searches. Do not repeatedly reboot while you are still establishing the boundary: a reboot can erase volatile state and obscure the original event.
2. When SSH fails, use an out-of-band console
SSH depends on a working network path, an accepting sshd, valid credentials, and a responsive kernel. Its failure is not proof that the operating system is powered off. Open the provider’s serial console, KVM, virtual console, or physical terminal. Confirm whether the machine is at a login prompt, stuck during boot, showing a kernel panic, or completely unresponsive.
Preserve evidence before rebooting
If the console responds, capture the screen, current time, boot messages, and any visible panic or OOM lines. Avoid clearing logs, restarting every service, or issuing a blind reset. First collect read-only state:
date -Is
uptime
systemctl --failed
systemctl status sshd --no-pager
ps aux --sort=-%mem | head -n 20
free -h
vmstat 1 5
df -hT
mountUse the service name used by the distribution; on some systems the SSH unit is ssh.service rather than sshd.service. If the system is so starved that commands hang, take console screenshots and note the symptoms before a controlled power action.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Distinguish a frozen host from a network-only outage
A console that accepts input while external probes time out suggests routing, firewall, interface, or service exposure problems. A console that stops responding, shows a kernel backtrace, or resets without a normal shutdown suggests a hard crash, hardware event, or virtual-machine failure. Provider dashboards may show host maintenance or power events; preserve those event times with your own observations.
Rank #2
- 【IP68 Waterproof Multifunctinal Smart Watch】Torjalph fitness activity tracker smart watch is IP68 waterproof technology ,you can wear it when swimming,handwashing,it's rain-proof which means that you perform multiple exercise modes outside without worrying about sweat or rain corroding your fitness smart watch,and this watch equipped with a 230 mAh large capacity battery magnetic charging ,and this sports activity tracker watch can be used 5-7 days If fully charged.
- 【Larger 1.83'' Full Screen Watch & Multi-dial Selection】1.83 inch full HD color touch screen smart watch for men / women provides excellent 240*284 resolution image quality ,This Bluetooth smart watch offers a better touch experience and remarkable immersion and ,more than 200 watch faces for you to choose from App"Da Fit ", you can change the dial style according to your mood to wear keeping the world on your wrist
- 【Accurate Activity Tracker Smart Watch 】sports watch activity tracker uses sweat and dust resistant breathable silicone strap, our Fitness Tracker Smart Watch for Android iOS Phones supports 12 sports modes,walking,running,bicycle,swimming,badminton,basketball,football,rope skipping. This bluetooth calls smart watch for men women is lightweight and stylish, bringing more convenient and relaxing sports experience while also providing more protection.
- 【Your 24 Hour Personal Daily Health Monitor】Fitness watch for women men full touch screen IP67 waterproof bluetooth smartwatch for Android iOS Phones with heart rate blood pressure monitor multifunction Andriod smart watch with steps counter sleep monitor sports fitness watch, monitoring blood oxygen saturation allows you to understand your physical condition in real time to track fitness progress ,to know your health status in time.
- 【More Functions Bluetooth Call Smart Watch】This fitness activity smart watch full screen for iPhone and Android phones includes a bunch of new features including:The bluetooth smart watch can hang up the phone, receive and display WhatsAPP /Text /Facebook /Twitter /Messenger reminder.Support sedentary reminder, alarm clock, stopwatch, timer,weather information,music control, remote camera control compatible with Android 5.0 and above or iOS 9.0 and above smartphones.
3. Inspect the current boot and the failing service
Once you have a shell, inspect the service state and the current boot’s journal before changing configuration. Start with the service that owns the failed endpoint, then inspect dependencies it reports as failed.
Use systemd status first
systemctl status nginx.service --no-pager
systemctl status apache2.service --no-pager
systemctl status your-app.service --no-pager
systemctl --failedReplace the examples with your actual units. Look for the exit code, the last start attempt, dependency failures, permission errors, missing files, and a configuration-test hint. A service can be active while the application behind it is unhealthy, so pair unit status with a local HTTP request and application logs.
Filter journald by boot, priority, service, and time
The following filters isolate the most useful evidence:
Recommended Free Tools
journalctl -b
journalctl -p err -b
journalctl -u your-app.service -b
journalctl --since '2026-10-15 12:00' --until '2026-10-15 13:00'journalctl -bshows messages from the current boot.journalctl -p err -blimits output to error and higher priorities for that boot.journalctl -u your-app.service -bfollows one service’s current-boot messages.--sinceand--untilnarrow a known incident window; use the outage time you recorded externally.
Read several lines before and after the first failure. A later “failed” message may be a consequence of an earlier disk, permission, dependency, or resource error. Export relevant excerpts to a separate incident file without truncating timestamps or boot identifiers.
Check kernel and hardware messages
journalctl -k -b
journalctl -k -p warning..alert -b
dmesg -T | tail -n 200Access to dmesg may require root privileges or be restricted by the distribution. Kernel messages containing OOM activity, I/O errors, filesystem complaints, machine-check reports, or a panic marker help separate an operating-system failure from an application failure.
4. Determine whether the machine rebooted
A server that is up now may have rebooted during the outage. Linux can only show earlier boots if journald was configured to retain them before the incident.
Enable persistent journal storage
On systems using systemd-journald, create the persistent directory and flush volatile entries:
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #3
- [Low Noise High Fidelity] New bee guitar cord with 22AWG oxygen-free-copper center conductors provides enhanced signal transfer, the PVC insulation material and the high-density stranded copper double shield provide noise rejection for quiet operation
- [Ideal for Choice] New bee 1/4 straight and right angle metal connectors, ensures a secure connection in any instrument, pedal, or amp jack. L shape great fro on stage guitar stands
- [Wide Compatible] Suitable for mixer, power amplifier, microphone, effector, decoder, equalizer, electoronic organ, electric guitar, electric bass and other electronic equipments. Widely used in the installation of smart home, karaoke speaker, professional stage audio, etc
- [Easy Disassemble] Making incidental repairs quick and easy! Very nice, flexible cable
- [Dont Hesite to Order] 24 hours customer services and professional technology team are standing by
sudo mkdir -p /var/log/journal
sudo systemd-tmpfiles --create --prefix /var/log/journal
sudo journalctl --flushConfirm that the journal configuration and retention policy leave enough disk space for the expected incident history. Verify persistence by checking that files exist under /var/log/journal and by testing a reboot during a maintenance window. Creating the directory after a crash cannot reconstruct earlier volatile entries.
Compare boot records
journalctl --list-boots
last -Fxn2 shutdown rebootIn journalctl --list-boots, boot 0 is the current boot, -1 is the immediately previous boot, and lower numbers are older boots. Query the previous boot directly:
journalctl -b -1
journalctl -b -1 -p err
journalctl -b -1 -kThe last query supplies recorded shutdown and reboot events. A clean shutdown entry supports an orderly restart; an abrupt end with no shutdown record is consistent with power loss, a reset, or a kernel-level failure, but it is not by itself proof of a panic. Correlate it with kernel messages, provider events, and crash-dump artifacts.
5. Check for memory and swap exhaustion
Exhausted memory can make a host appear dead because reclaim and I/O consume all CPU time, or because the kernel’s out-of-memory (OOM) handler terminates critical processes. Capture resource state before adding swap, killing processes, or restarting services.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Capture a resource snapshot
free -h
swapon --show
vmstat 1 10
ps aux --sort=-%mem | head -n 30
ps aux --sort=-%cpu | head -n 30
df -hT
df -ihIn free -h, compare available memory with swap used; in vmstat, sustained swap-in and swap-out activity indicates pressure. Check both block-space and inode exhaustion because a full filesystem can prevent services from writing logs, sockets, PID files, or application data.
Find OOM evidence
journalctl -k -b | grep -i -E 'out of memory|oom|killed process'
journalctl -p err -b | grep -i -E 'memory|oom|killed'Identify the process named in the kernel message and determine whether a leak, traffic spike, unbounded worker pool, or a mis-sized cache caused growth. A practical remedy may be reducing workload, correcting the leak, limiting workers, or adding RAM. Adding swap can provide emergency headroom but is slower than RAM and may turn an abrupt kill into severe latency; size and monitor it deliberately rather than treating it as a universal fix.
6. Recover a host that fails during normal startup
If the kernel boots but the machine hangs while services start, use the provider console to enter a rescue target. At the boot loader, append systemd.unit=rescue.target (also known as runlevel 1) to the kernel command line, then continue booting. The exact boot-loader editing key varies by distribution.
Rank #4
- Digital Family Calendar: BSIMB digital calendar wall touch screen puts an end to your chaotic life. Designed for busy families, the digital calendar keeps you organised and in step with your family. Assign different colours and tasks to each family member to make family scheduling easier and more intuitive.
- Easy Setup and Auto-Sync: Plug in, connect to WLAN and auto-syn your electric calendar in minutes. Digital wall calendar compatible with Google calendar, iCloud calendar, Outlook calendar, Cozi calendar and Yahoo calendar. Add events directly on the device or through the ecalendar app. Electronic calendar stays current whether you're at work or traveling—no manual updates needed.
- All-In-One Home Assistant: Smart calendar builds habits and encourage independence with personalised chores. Display daily meal plan to reduce the daily ‘What's for dinner?’ stress. To-do lists keep track of everything you want. Turning chores into valuable achievements with stars and rewards, making life easier and more enjoyable.
- Manage Anytime Anywhere: BSIMB electronic calendar for family stay connected with the powerful ecalendar app. Access and manage your digital wall calendar anytime, anywhere with free features like stars & rewards, photo screensaver, chore charts, meals plans, to-do lists etc. Electronic planner paid features like Magic Import tailor advanced tools for users with special needs.
- Stay in Touch with Loved Ones: Share photos or videos from your phone to electronic wall calendar with app, instantly update your life moment, and keep in touch with your family, especially those who are far away; you can also surprise your smart calendar digital wall by sending him a wish card; it's also great for keeping tech-illiterate elders in touch with the youngsters.
Work methodically in rescue mode
- Record the boot messages and identify the unit at which progress stops.
- Inspect failed units with
systemctl --failedand read their current-boot logs usingjournalctl -b. - Check mounts, disk space, and filesystem availability before editing service files.
- Validate configuration with the application’s own test command, such as the web server’s configuration check, before restarting it.
- Repair the specific dependency, credential, path, package, or configuration error; do not disable unrelated security or storage services merely to make boot continue.
- Exit rescue mode or reboot to the normal target only after the failing component has a verified configuration and a recovery path.
If the problem began after a package, kernel, or deployment change, preserve the installed versions and change times in the incident record. Roll back only with a tested procedure and a backup or snapshot.
7. Investigate a kernel panic or hard crash
A kernel crash dump is a portion of volatile RAM copied to disk when kernel execution is disrupted. Documented triggers include a kernel panic, non-maskable interrupt, machine-check exception, hardware failure, or manual intervention. The dump is often the strongest evidence when the normal journal ends abruptly.
Check kdump readiness
On Ubuntu systems using the documented tooling, inspect:
kdump-config show
systemctl status kdump-tools
cat /proc/cmdline
ls -lah /var/crashLook for a reserved crashkernel region in the kernel command line or boot configuration, a running crash-dump service, and a dump destination with sufficient space. Ubuntu stores crash artifacts under /var/crash when configured to do so. An “not ready” state can result from insufficient reserved memory; increasing the reservation and rebooting is a planned change, not an incident-time experiment.
Read the dump without destroying it
Preserve the vmcore and its metadata before opening or transferring it. Match the dump to the exact kernel build and collect the corresponding symbols or debug packages. A crash-analysis tool can then identify the panic trace, tasks, and modules involved. Keep the original file read-only and record its path, size, checksum, host, and boot time.
Do not trigger a crash over SSH
SysRq crash testing intentionally reboots the machine and can cause data loss under load. It also drops the SSH connection, so testing requires an approved maintenance window, a local or serial console, current backups, and a recovery plan. Never send echo c > /proc/sysrq-trigger to a production host merely to see whether kdump works. Validate crash capture on an idle or lightly loaded test system first, following the distribution’s documented procedure.
8. Choose recovery actions by failure layer
The safest response depends on what you can prove. The table below keeps evidence, access, and risk visible while you decide.
| Observed condition | Primary access | First evidence | Typical immediate action | Main risk |
|---|---|---|---|---|
| DNS answer missing or incorrect | Independent resolver and DNS control plane | DNS records, TTL, resolver responses | Correct the record or investigate the authoritative provider | Changing records can prolong propagation or hide the original error |
| DNS works; TCP or HTTP fails, SSH works | SSH or console | Socket listeners, firewall state, service journal | Repair the affected service or network rule after configuration validation | Restarting can erase the first failure context |
| SSH and HTTP fail; console responds | Serial, KVM, or provider console | Current boot journal, kernel messages, memory, disk | Capture state, then repair resource, mount, dependency, or network issues | Heavy commands or broad restarts can worsen a saturated host |
| Console shows rescue loop or startup hang | Console and rescue target | Boot messages, failed units, dependency ordering | Boot to systemd.unit=rescue.target, validate and repair the named component |
Disabling dependencies may create a less visible failure |
| Console resets or shows panic evidence | Console plus crash-dump storage | vmcore, panic trace, previous-boot journal, provider events | Preserve artifacts, restore from a known-good state, then analyze the exact kernel and hardware context | A reset or dump misconfiguration can destroy volatile evidence |
Time to restore and data-loss exposure vary with the service, storage, provider, and backup design. Treat a reboot as a recovery action only after you have collected the evidence that can still be collected.
9. Restore service without losing the investigation
Apply the smallest reversible change
Prefer a configuration correction, dependency repair, capacity limit, or failover over a broad package purge. Save the original file and note every command, timestamp, and operator. For a web service, run its syntax check, restart only that unit, and test locally before exposing it to external traffic.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsVerify each layer after recovery
- Confirm the host has stable uptime and no continuing kernel, OOM, filesystem, or hardware errors.
- Confirm the expected process is listening with
ss -ltnp. - Request the service locally and through the public hostname, checking status code and latency.
- Test SSH, application dependencies, background workers, and scheduled jobs.
- Watch the journal and resource metrics for a sustained period instead of declaring success after one request.
If the root cause is not established, document the service as restored but the incident as unresolved. A temporary restart can return traffic while leaving a leak, failing disk, bad kernel, or broken deployment in place.
10. Build an incident record that supports root-cause analysis
Keep one chronological record containing:
- The first observed outage time, monitoring alerts, DNS results, TCP outcomes, HTTP status codes, and test locations.
- Provider status or maintenance events and the console access path used.
- Boot IDs from
journalctl --list-boots, relevantjournalctlexcerpts, service status, and kernel messages. - Memory, swap, CPU, filesystem, inode, mount, and process snapshots taken before changes.
- Recent deployments, package updates, kernel changes, configuration edits, and infrastructure alterations.
- Any OOM lines, panic traces, vmcore path, dump metadata, and checksums.
- Every remediation command, its result, the time service was restored, and post-recovery verification.
Preserve the earliest OOM or panic lines with surrounding context. An isolated line rarely establishes causality; the sequence across boots, services, resources, and provider events is what distinguishes a trigger from a consequence.
11. Prevent the next opaque outage
Make logs survive a reboot
Enable persistent journald storage, set a retention policy that fits the disk, and periodically verify that previous boots remain queryable. Alert when the journal filesystem is near capacity so logging does not silently stop.
Monitor the failure signals that matter
- Memory availability, swap-in and swap-out activity, and OOM-killer events.
- Filesystem and inode utilization, I/O errors, and mount failures.
- HTTP and TCP health from outside the host, plus SSH and critical dependency checks.
- Service state changes, repeated restarts, failed units, and reboot events.
- Kernel warnings, machine-check reports, and crash-dump readiness.
Keep recovery access and copies ready
- Document serial, KVM, or provider-console access, ownership of credentials, and an emergency contact path.
- Maintain tested backups and provider snapshots; know the restore time and the data-loss point for each.
- Configure kdump where the distribution and hardware support it, reserve adequate crashkernel memory, and allocate dump storage.
- Practice a log-collection and restoration runbook during a maintenance window, including a console-only scenario.
Run a harmless rehearsal: verify that an alert fires, a previous boot is visible, the console can be opened, and a backup can be restored. Do not substitute an unapproved production crash test for that rehearsal.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
