====== Ubuntu Server Update Procedure ======

This is the normal update procedure for the Ubuntu server.

The goal is to keep the system updated while avoiding problems caused by kernel or NVIDIA driver updates that require a reboot.

===== 1. Refresh Package Information =====

Run:

<code bash>
sudo apt -o Acquire::ForceIPv4=true update
</code>

Using ''Acquire::ForceIPv4=true'' avoids the Ubuntu mirror connectivity problem that can occur when the server attempts IPv6 connections that are not available.

===== 2. Review Available Updates =====

Run:

<code bash>
apt list --upgradable
</code>

Review the list before installing anything.

Pay particular attention to updates involving:

<code>
linux-image
linux-headers
nvidia-driver
nvidia-utils
libnvidia
nvidia-kernel
</code>

Kernel or NVIDIA updates may require a reboot.

===== 3. Install Updates =====

Run:

<code bash>
sudo apt -o Acquire::ForceIPv4=true upgrade
</code>

When prompted:

<code>
Continue? [Y/n]
</code>

Enter:

<code>
y
</code>

Ubuntu may leave some packages uninstalled because of "phasing." This is normal.

Do not force phased updates unless there is a specific reason to do so.

===== 4. Check Whether a Reboot Is Required =====

After the update finishes, run:

<code bash>
if [ -f /var/run/reboot-required ]; then
  echo "REBOOT REQUIRED"
  cat /var/run/reboot-required.pkgs 2>/dev/null
else
  echo "No reboot required"
fi
</code>

If a reboot is required, the output may also list the package or packages responsible.

===== 5. Reboot If Required =====

If Ubuntu reports that a reboot is required:

<code bash>
sudo reboot
</code>

After the server comes back online, reconnect with SSH.

===== NVIDIA / Plex Warning =====

If NVIDIA libraries are updated without rebooting, the loaded NVIDIA kernel module may no longer match the installed NVIDIA libraries.

This can cause Plex Media Server to crash immediately at startup.

A symptom of this condition is:

<code>
Failed to initialize NVML: Driver/library version mismatch
</code>

If this appears after an NVIDIA update, reboot the server before troubleshooting Plex.

After rebooting, verify the NVIDIA driver with:

<code bash>
nvidia-smi
</code>

A normal ''nvidia-smi'' display indicates that the NVIDIA driver and libraries are communicating correctly.

===== 6. Check Drive Mounts =====

After updates or a reboot, run mount check for the drives:

<code bash>
for m in /mnt/media{1..8} /mnt/internal; do
  if mountpoint -q "$m"; then
    echo "OK      $m"
  else
    echo "MISSING $m"
  fi
done </code>

===== 7. Check Server Services =====

After updates or a reboot, run the server health check:

<code bash>
echo "=== Native Services ==="
for s in plexmediaserver transmission-daemon jdownloader; do
    printf "%-22s %s\n" "$s" "$(systemctl is-active "$s" 2>/dev/null)"
done

echo
echo "=== Docker Services ==="
for c in nginx-proxy-manager newsblur_web newsblur_celery newsblur_nginx newsblur_haproxy; do
    printf "%-22s " "$c"
    docker inspect -f '{{.State.Status}}' "$c" 2>/dev/null || echo "not found"
done </code>

Expected results:

<code>
=== Native Services ===
plexmediaserver        active
transmission-daemon    active
jdownloader            active

=== Docker Services ===
nginx-proxy-manager    running
newsblur_web           running
newsblur_celery        running
newsblur_nginx         running
newsblur_haproxy       running </code>

===== Recommended Update Commands =====

For normal updates, use:

<code bash>
sudo apt -o Acquire::ForceIPv4=true update
apt list --upgradable
sudo apt -o Acquire::ForceIPv4=true upgrade
</code>

For this server, normally avoid using:

<code>
apt full-upgrade
apt dist-upgrade
</code>

unless there is a specific reason to do so.

Also do not force phased updates unless necessary.

===== Quick Summary =====

The normal routine is:

* Refresh package information.
* Review available updates.
* Install normal eligible updates.
* Check whether Ubuntu requires a reboot.
* Reboot when required, especially after kernel or NVIDIA updates.
* Run ''nvidia-smi'' after NVIDIA-related updates.
* Run the server health check after rebooting.