Quick Start Checklist

If you are new to Karios, start here first. Use this checklist before starting full deployment. Typical total time: 15-30 minutes for single-node readiness checks.

Recommended first-time user path:

  1. Complete this checklist.

  2. Read Getting Started for context and platform overview.

  3. Continue to ISO Install for the 9-step graphical installer and platform image delivery.

  4. After the installer completes and the node reboots, open Control Center, the Karios web UI, at the final HTTPS URL shown on the console.

Terms used in this checklist:

  • Control Center: the main Karios web UI you open in a browser after the installer completes and the node reboots.

  • Platform image (the template): the Karios .qcow2 image (approximately 5 GB) you deliver from a second machine when the install pauses partway through. It is bundled with the ISO in the same .zip download — you do not fetch it separately.

  • Graphical Installer vs Control Center: the installer runs a 9-step wizard on the server console or remote KVM session and then reboots the node directly into first boot; normal platform management then moves to the final HTTPS Karios UI URL.

  • AIO (all-in-one): a node running every role on one chassis — control plane, storage, and compute. Needs 2+ disks (one or more for the OS, at least one left empty for the data pool). Fault-tolerant HA needs 3 or more nodes.

  • Management-only: a node running only control-plane services and databases — no local VMs or storage workload. A single disk is enough. Typically deployed as three nodes for a highly-available control plane.

  • BMC/IPMI: the server’s out-of-band remote management controller, such as IPMI, iDRAC, or iLO.

  • VT-x/AMD-V: CPU virtualization features that must be enabled in BIOS for VM workloads.

  • CIDR: network range notation such as 10.0.1.0/24.

  • Ceph: the distributed storage layer. On an AIO node it claims the disk you leave empty.

  • Instance profile: a reusable VM sizing definition (CPU/RAM). At least one must exist before you can create a VM.

  • OOB: out-of-band management network used for hardware management access. Optional, and not currently used by the installer.

Note

If these terms are unfamiliar, finish this checklist first, then read Getting Started before continuing to ISO installation.

Hardware Validation

Estimated time: 3-5 minutes

  • [ ] Management node: 16+ CPU cores

  • [ ] Management node: 16+ GB RAM installed

  • [ ] Disks: AIO node 2+ (at least one left completely empty for Ceph); Management-only node 1+

  • [ ] Each disk counted above is internally attached and at least 32 GiB — USB/removable, SAN/iSCSI/FC/FCoE, and the install medium do not count

  • [ ] Disk controller set to HBA / JBOD / pass-through mode (the installer refuses hardware-RAID volumes). Changing controller mode destroys any existing hardware-RAID volume — do this before there is anything on the machine worth keeping

  • [ ] BIOS/UEFI: firmware set to UEFI (Legacy/CSM boot is refused)

  • [ ] BIOS/UEFI: Secure Boot disabled (or plan to enroll the Karios key from the local/BMC console when prompted)

  • [ ] BIOS/UEFI: VT-x/AMD-V available for virtualization workloads

  • [ ] Node role decided: AIO (all-in-one, 2+ disks) or Management-only (single disk)

  • [ ] BMC/IPMI reachable with valid credentials

  • [ ] If planning agent / compute nodes: 16+ CPU cores, 16+ GB RAM, KVM acceleration, and 3+ disks excluding the OS disk

  • [ ] If planning agent / compute nodes: OS disk is Tier 1 (10+ DWPD) or Tier 2 (3-10 DWPD), and remaining disks include at least one Tier 1, Tier 2, and Tier 3 (1-3 DWPD)

  • [ ] Hardware vendor confirmed as Supermicro, Dell, or HP

HA Preflight (Multi-Node Deployments)

Estimated time: 5-10 minutes (skip for single-node labs)

  • [ ] Plan at least 3 hosts for resilient cluster behavior

  • [ ] Confirm NTP and DNS consistency across all planned hosts

  • [ ] Confirm management and storage networks are reachable from all hosts

  • [ ] Confirm each host has compatible CPU virtualization features and similar baseline capacity

  • [ ] Confirm failure-domain plan (rack/power/network) before production onboarding

Platform Image Delivery

Estimated time: download-bound — start this first, it runs while you do the hardware work.

  • [ ] Karios installer .zip downloaded from https://download.karios.com/installer/karios-installer.zip and extracted — the one download contains both the ISO and the platform image (the template)

  • [ ] Image is on a second machine that is on, or routable to, the management network (the node has no address until the installer’s Network step)

  • [ ] That machine has a browser — the install pauses partway through and waits for the image

  • [ ] Platform image matches the ISO version being installed

Network Planning

Estimated time: 3-5 minutes

  • [ ] Management CIDR planned (example: 10.0.1.0/24)

  • [ ] Storage CIDR planned (example: 10.0.2.0/24)

  • [ ] Public CIDR planned (example: 10.0.3.0/24)

  • [ ] Guest CIDR planned (example: 10.0.4.0/24)

  • [ ] BMC CIDR planned (the OOB installer role itself is unused) (example: 192.168.100.0/24)

  • [ ] All 4 required VLAN IDs planned and unique (Management, Storage, Public, Guest)

  • [ ] DNS servers identified

  • [ ] NTP servers identified

  • [ ] Node’s own static management IP chosen and confirmed unused

  • [ ] Default gateway for the management network confirmed

  • [ ] Hostname chosen, and the DNS record for it created if operators will reach Control Center by name

  • [ ] NIC wiring decided: one NIC (Management untagged + Storage/Public/Guest tagged on the same trunk) or two NICs (Management untagged, second NIC trunks the other three)

  • [ ] Switch ports configured to trunk the planned VLANs before the install

  • [ ] No overlap across Management/Storage/Public/Guest CIDRs (and OOB if used)

  • [ ] Management usable IP range is greater than 20 IPs

  • [ ] Storage usable IP range is greater than 20 IPs and not smaller than Management

  • [ ] BMC network DHCP disabled

Post-Install Validation

Note

The sections above are the pre-install gate. Everything from here on is checked after the install completes and the node reboots.

Estimated time: 3-5 minutes

  • [ ] Control Center accessible at the final HTTPS URL shown on the console (FQDN or management IP)

  • [ ] Admin login successful

  • [ ] At least one instance profile visible

  • [ ] At least one guest network available (the default guest network created during installation is acceptable)

  • [ ] Primary storage shows Up state

  • [ ] At least one VM template or boot image in Ready state (a VM template is a guest OS image for provisioning VMs — not the platform image you delivered during the install)

If instance profile is missing:

  1. Check Compute Policies and verify the Instance Profiles dashboard has at least one active row.

  2. If no profile exists, create one from Instance Profiles.

  3. Re-check VM/Kubernetes create forms after adding the profile.

First VM Ready State

Estimated time: 2-5 minutes (template path)

  • [ ] Test VM provisioned successfully

  • [ ] VM reaches Running state

  • [ ] VM console accessible

  • [ ] VM network connectivity validated