20250514-notes

projects/proxmox/thomas-krenn-schlulung-part1/proxmox-notes-part1.html

@@ -0,0 +1,102 @@
+<h1
+id="proxmox-virtual-environment-and-clustering-with-zfs-and-ceph---introductory-course-notes">Proxmox
+Virtual Environment and Clustering with ZFS and Ceph - Introductory
+Course Notes</h1>
+<h2 id="introduction">Introduction</h2>
+<h3 id="basics">Basics</h3>
+<ul>
+<li>Debian stable + Ubuntu Kernel</li>
+<li>Virtualization with KVM</li>
+<li>LXC Containers</li>
+<li>Supports ZFS and Ceph</li>
+</ul>
+<h3 id="features">Features</h3>
+<ul>
+<li>Snapshots</li>
+<li>KVM Virtual Machines (Windows, Linux, BSD)</li>
+<li>LXC Containers</li>
+<li>High Availability (HA) Clustering</li>
+<li>Live Migration</li>
+<li>Flexible Storage Options</li>
+<li>GUI Management</li>
+<li>Proxmox Data Center Manager</li>
+<li>Proxmox Backup Server</li>
+</ul>
+<h2 id="virtualization-stack">Virtualization Stack</h2>
+<figure>
+<img src="/files/proxmox/2025-05-14.png" alt="Virtualization Stack" />
+<figcaption aria-hidden="true">Virtualization Stack</figcaption>
+</figure>
+<h2 id="expansion-options">Expansion Options</h2>
+<h3 id="single-node">Single Node</h3>
+<ul>
+<li><strong>Storage Pool:</strong>
+<ul>
+<li>RAID Controller with LVM – Ceph and ZFS don’t support a RAID
+controller; if a RAID controller is used, change to the mode and use
+ZFS.</li>
+</ul></li>
+</ul>
+<h3 id="clustering-without-ha">Clustering Without HA</h3>
+<ul>
+<li>Several Storage Pools</li>
+<li>Pools are not shared – individual</li>
+</ul>
+<h3 id="zfs-cluster-with-asynchronous-replication">ZFS Cluster with
+Asynchronous Replication</h3>
+<ul>
+<li>At least two pools</li>
+<li>A quorum device between them replicates the configuration between
+the nodes.</li>
+</ul>
+<h3 id="ceph-cluster">Ceph Cluster</h3>
+<ul>
+<li>At least 3 nodes</li>
+<li>Pool is combined and shared</li>
+</ul>
+<h2 id="installation">Installation</h2>
+<ul>
+<li>It’s possible to provide custom TOML files to install Proxmox
+automatically.</li>
+<li>ZFS RAID1 for the boot drive.</li>
+<li>Define the correct hostname; it’s not easy to change later.</li>
+<li>Modify the update repositories as a first step! And update/upgrade
+the system.</li>
+<li>Hardware: Always better to have as many interface ports as possible.
+<ul>
+<li>2x 1 Gbit for Management</li>
+<li>2x 10 Gbit for VMs (redundant, bonded via LACP – requires stacked
+switches or backup mode)</li>
+<li>2x 25 Gbit for Ceph storage clustering</li>
+<li>Single node: at least 4 ports; for clustering: at least 6
+ports.</li>
+</ul></li>
+<li>Remove the IP address of the virtual bridge and set it directly on
+the physical interface. This prevents VMs from seeing the Proxmox
+IP.</li>
+<li>Bond ports for VMs (ideally two 10 Gbit ports) and set the virtual
+bridge on <code>bond0</code> so it can be used by VMs (set ‘VLAN aware’
+at the bridge).</li>
+<li>It’s possible to create a bond for the web UI port.</li>
+</ul>
+<h2 id="vm-creation">VM Creation</h2>
+<h2 id="todo">Todo</h2>
+<ul class="task-list">
+<li><label><input type="checkbox" />Join the forum</label></li>
+<li><label><input type="checkbox" />Research about initial TOML file for
+Proxmox for unattended configuration of PVE - ‘proxmox
+autoinstaller’</label></li>
+<li><label><input type="checkbox" />Test ZRAID1 on two SSDs at home with
+a ‘turtle’ Proxmox host</label></li>
+</ul>
+<h2 id="support-subscriptions">Support Subscriptions</h2>
+<h3 id="basic">Basic</h3>
+<p>1 CPU per year / 2 CPUs per year * Access to enterprise repositories
+* Stable software updates * Support via the customer portal * Three
+support tickets * Response time: 1 business day</p>
+<h3 id="standard">Standard</h3>
+<p>1 CPU per year / 2 CPUs per year * All Basic features</p>
+<p><strong>Additionally:</strong> * 7 support tickets (total 10) *
+Remote support (via SSH) * Offline subscription key activation</p>
+<h3 id="premium">Premium</h3>
+<p>1 CPU per year / 2 CPUs per year * All Standard features</p>