36 lines
1.4 KiB
Markdown
36 lines
1.4 KiB
Markdown
---
|
|
id: f79b6b07-a0ec-4baa-9e1b-4732d104056a
|
|
type: solution
|
|
title: "Proxmox 7 to 8 upgrade completed successfully"
|
|
tags: [proxmox, homelab, upgrade, infrastructure]
|
|
importance: 0.9
|
|
confidence: 0.8
|
|
created: "2026-02-19T17:05:30.467207+00:00"
|
|
updated: "2026-02-19T17:05:30.467207+00:00"
|
|
---
|
|
|
|
## Proxmox PVE 7.4-20 to 8.4.16 Upgrade (2026-02-19)
|
|
|
|
Successfully upgraded Proxmox from PVE 7.4-20 (kernel 5.13.19) to PVE 8.4.16 (kernel 6.8.12-18-pve).
|
|
|
|
## Procedure
|
|
1. Backed up all 13 production guests to TrueNAS (home-truenas PVE storage at /mnt/pve/home-truenas)
|
|
2. Ran pve7to8 --full (23 passed, 0 failures)
|
|
3. Switched local DNS to 1.1.1.1/8.8.8.8 (Pi-holes on Proxmox guests)
|
|
4. Shut down all guests (some needed --forceStop)
|
|
5. Backed up /etc/apt/sources.list and sources.list.d/
|
|
6. Changed repos: bullseye -> bookworm
|
|
7. DEBIAN_FRONTEND=noninteractive apt dist-upgrade -y -o Dpkg::Options::='--force-confnew'
|
|
8. Rebooted (~4 min to come back)
|
|
9. Started guests in dependency order
|
|
10. Validated all services
|
|
|
|
## Key Issues
|
|
- I/O storm: all guests booting simultaneously caused ~50% I/O delay for minutes
|
|
- Pi-hole 1 Docker container needed restart (UDP DNS failed, TCP worked)
|
|
- Home Assistant got new DHCP IP (10.10.0.215 instead of previous)
|
|
- VMs 105/112 ignored ACPI shutdown, needed --forceStop
|
|
- Several guests had onboot=1, auto-started before manual ordering
|
|
|
|
## Phase 2 (PVE 8->9) pending after 1-2 week stabilization period.
|