|
|
|
@@ -16,9 +16,9 @@
|
|
|
|
|
"items": [
|
|
|
|
|
"<strong>Verified backups</strong> of every VM and CT (PBS, vzdump, ZFS snapshots — and verify they restore).",
|
|
|
|
|
"<strong>Console / IPMI / iKVM access</strong> to the host for when SSH drops mid-upgrade.",
|
|
|
|
|
"The host already up-to-date on PVE 8.4.x (the script enforces this).",
|
|
|
|
|
"The host on the latest PVE 8 patch level. The script updates PVE 8 first; it does not check for a specific 8.4.x version.",
|
|
|
|
|
"Ceph (if present) on version 19.x Squid.",
|
|
|
|
|
"At least <strong>5 GB free</strong> on the root filesystem."
|
|
|
|
|
"At least <strong>5 GB free</strong> on the root filesystem, as the Proxmox guide recommends. The script itself only checks for 1 GB in /var/cache/apt/archives."
|
|
|
|
|
]
|
|
|
|
|
},
|
|
|
|
|
"modeMenu": {
|
|
|
|
@@ -63,13 +63,13 @@
|
|
|
|
|
},
|
|
|
|
|
"webTerminal": {
|
|
|
|
|
"heading": "Why the web terminal is blocked",
|
|
|
|
|
"body": "All three upgrade modes detect <code>termproxy</code> / <code>vncshell</code> in the parent process chain and refuse to run. Reason: the Proxmox web terminal is served by <code>pveproxy</code>, which is one of the services that gets upgraded mid-flow. As soon as <code>pveproxy</code> restarts, your terminal session dies — and so does the upgrade script that was running inside it. The host is left half-upgraded, no good way to resume.",
|
|
|
|
|
"body": "The Automatic and Interactive modes detect <code>termproxy</code> / <code>vncshell</code> in the parent process chain and refuse to run, unless the session comes over SSH or <code>PVE_ALLOW_WEBTERM=1</code> is set. The pre-check and the manual guide change nothing and run anywhere. Reason: the Proxmox web terminal is served by <code>pveproxy</code>, which is one of the services that gets upgraded mid-flow. As soon as <code>pveproxy</code> restarts, your terminal session dies — and so does the upgrade script that was running inside it. The host is left half-upgraded, no good way to resume.",
|
|
|
|
|
"warningTitle": "Use SSH or physical / IPMI / iKVM console",
|
|
|
|
|
"warningBody": "SSH is the most common choice. Open it inside <code>tmux</code> or <code>screen</code> so the upgrade survives a brief network blip. The very first thing the script does is print the tmux/screen reminder if it detects you're in a plain SSH shell."
|
|
|
|
|
"warningBody": "SSH is the most common choice. Open it inside <code>tmux</code> or <code>screen</code> so the upgrade survives a brief network blip. The upgrade confirmation dialog reminds you to run it from the console or from SSH inside tmux/screen."
|
|
|
|
|
},
|
|
|
|
|
"auto": {
|
|
|
|
|
"heading": "Automatic / Unattended mode",
|
|
|
|
|
"intro": "Package and configuration-file prompts use the defaults in the table below, and a 401 from the enterprise repository switches to no-subscription without asking. The upgrade confirmation, the PVE 8 update summary, any pre-check repairs and the reboot prompt are still shown.",
|
|
|
|
|
"intro": "Package and configuration-file prompts use the defaults in the table below. The upgrade confirmation, the PVE 8 update summary, any pre-check repairs and the reboot prompt are still shown.",
|
|
|
|
|
"behaviourHeading": "Auto-mode-specific behaviour",
|
|
|
|
|
"behaviourTable": {
|
|
|
|
|
"settingHeader": "Setting",
|
|
|
|
@@ -94,8 +94,8 @@
|
|
|
|
|
]
|
|
|
|
|
},
|
|
|
|
|
"flowHeading": "Auto-mode flow",
|
|
|
|
|
"flowIntro": "Six sequential phases. Phase 2 updates the PVE 8 packages and phase 3 writes the PVE 9 repository configuration. A failure stops the flow without rolling those changes back; if the PVE 8 update fails or is cancelled, the repositories are left unchanged. Phase 5 runs the PVE 9 <code>dist-upgrade</code>.",
|
|
|
|
|
"flowDiagram": "1. CHECKS\nBlock termproxy / vncshell; confirm the upgrade; reject PVE 9.\n\n2. PREPARE PVE 8\nscripts/global/update-pve8.sh\nUpdate PVE 8 packages; stop if the update fails or is cancelled.\npve8to9 --full (may offer repairs).\nCheck space and connectivity; Ceph 19.x Squid on hyper-converged nodes.\n\n3. MIGRATE REPOSITORIES\nBookworm → Trixie; write deb822 .sources files.\nComment legacy .list files; apt-get update.\n\n4. VALIDATE THE PVE 9 CANDIDATE\napt-cache policy proxmox-ve\napt-get -s dist-upgrade\nStop if the candidate is not 9.x or proxmox-ve would be removed.\n\n5. UPGRADE\napt-get -y -o Dpkg::Options::='--force-confdef' -o Dpkg::Options::='--force-confold' dist-upgrade\nNo subscription: disable pve-enterprise.sources.\nLog: /var/log/pve8-a-pve9-<timestamp>.log\n\n6. VERIFY AND REBOOT\nEFI: apt install grub-efi-amd64\npve8to9 --full; review results and the reboot prompt.",
|
|
|
|
|
"flowIntro": "Six sequential phases. Phase 2 switches the PVE 8 repositories to no-subscription (enterprise and Ceph lists disabled, <code>/etc/apt/sources.list</code> rewritten) before it shows the update summary, then updates the packages; phase 3 writes the PVE 9 repository configuration. A failure stops the flow without rolling those changes back; if the PVE 8 update fails or is cancelled, the PVE 9 migration does not start, but the PVE 8 repository changes stay. Phase 5 runs the PVE 9 <code>dist-upgrade</code>.",
|
|
|
|
|
"flowDiagram": "1. CHECKS\nBlock termproxy / vncshell; confirm the upgrade; reject PVE 9.\n\n2. PREPARE PVE 8\nscripts/global/update-pve8.sh\nSwitch PVE 8 repositories to no-subscription and rewrite sources.list;\nupdate PVE 8 packages; stop if the update fails or is cancelled.\npve8to9 --full (may offer repairs).\nCheck space and connectivity; Ceph 19.x Squid on hyper-converged nodes.\n\n3. MIGRATE REPOSITORIES\nBookworm → Trixie; write deb822 .sources files.\nComment legacy .list files; apt-get update.\n\n4. VALIDATE THE PVE 9 CANDIDATE\napt-cache policy proxmox-ve\napt-get -s dist-upgrade\nStop if the candidate is not 9.x or proxmox-ve would be removed.\n\n5. UPGRADE\napt-get -y -o Dpkg::Options::='--force-confdef' -o Dpkg::Options::='--force-confold' dist-upgrade\nNo subscription: disable pve-enterprise.sources.\nLog: /var/log/pve8-a-pve9-<timestamp>.log\n\n6. VERIFY AND REBOOT\nEFI: apt install grub-efi-amd64\npve8to9 --full; review results and the reboot prompt.",
|
|
|
|
|
"preflightHeading": "Pre-flight: what gets checked",
|
|
|
|
|
"preflightTable": {
|
|
|
|
|
"checkHeader": "Check",
|
|
|
|
@@ -158,7 +158,7 @@
|
|
|
|
|
},
|
|
|
|
|
"manual": {
|
|
|
|
|
"heading": "Manual upgrade procedure",
|
|
|
|
|
"intro": "This is what the <strong>Manual upgrade guide</strong> menu option displays — a faithful mirror of <wikilink>the official Proxmox 8-to-9 upgrade wiki</wikilink> organised in 4 phases. The script <em>shows</em> these commands; you type them yourself in a separate root shell. No automation — full visibility.",
|
|
|
|
|
"intro": "This follows what the <strong>Manual upgrade guide</strong> menu option displays, based on <wikilink>the official Proxmox 8-to-9 upgrade wiki</wikilink> and organised in 4 phases. The dpkg answer table adds the wiki's caveats for sshd_config and chrony.conf, which the menu shows as a plain Y. The script <em>shows</em> these commands; you type them yourself in a separate root shell. No automation — full visibility.",
|
|
|
|
|
"rootCalloutTitle": "All commands need root",
|
|
|
|
|
"rootCalloutBody": "Use a root shell for these commands, including redirections and here-documents. If you use <code>sudo</code>, ensure the entire command runs with root privileges.",
|
|
|
|
|
"phase1Heading": "Phase 1 — Preparation",
|
|
|
|
@@ -230,7 +230,7 @@
|
|
|
|
|
"clusterCeph": {
|
|
|
|
|
"heading": "Cluster & Ceph",
|
|
|
|
|
"clusterCalloutTitle": "Cluster: upgrade nodes one at a time",
|
|
|
|
|
"clusterCalloutBody": "Detected via <code>pvecm status</code>. <strong>Do not</strong> upgrade more than one node at a time. Move guests off the node, upgrade, reboot and verify cluster membership before proceeding. During the upgrade, live migration from an older PVE version to a newer one works; the reverse direction is generally unsupported. HA groups migrate automatically to HA rules once all nodes run PVE 9.",
|
|
|
|
|
"clusterCalloutBody": "<strong>Do not</strong> upgrade more than one node at a time. Move guests off the node, upgrade, reboot and verify cluster membership before proceeding. During the upgrade, live migration from an older PVE version to a newer one works; the reverse direction is generally unsupported. HA groups migrate automatically to HA rules once all nodes run PVE 9.",
|
|
|
|
|
"cephCalloutTitle": "Ceph: upgrade Ceph FIRST, then Proxmox",
|
|
|
|
|
"cephCalloutBody": "Ceph must already run 19.x Squid before upgrading Proxmox. Follow the official sequence: Quincy 17 → Reef 18 → Squid 19, updating packages and restarting Ceph daemons in the documented rolling order. Verify <code>ceph --version</code> before upgrading each Proxmox node. Follow the full procedure: <cephlink>Ceph Reef to Squid upgrade wiki</cephlink>."
|
|
|
|
|
},
|
|
|
|
@@ -273,7 +273,7 @@
|
|
|
|
|
},
|
|
|
|
|
"files": {
|
|
|
|
|
"heading": "Files involved",
|
|
|
|
|
"code": "scripts/utilities/upgrade_pve8_to_pve9.sh # main upgrade orchestrator\nscripts/utilities/pve8to9_check.sh # pre-check (separate menu option)\nscripts/utilities/proxmox-upgrade-pve8-to-pve9-manual-guide.sh # 17-step runbook\nscripts/global/update-pve8.sh # PVE 8 worker (final patches)\n/etc/apt/sources.list # bookworm → trixie sed\n/etc/apt/sources.list.d/proxmox.sources # created (deb822 PVE 9)\n/etc/apt/sources.list.d/debian.sources # created (deb822 trixie)\n/etc/apt/sources.list.d/ceph.sources # created if Ceph\n/etc/apt/sources.list.d/*.list # legacy files commented out / removed\n/var/log/pve8-a-pve9-<timestamp>.log # full upgrade log"
|
|
|
|
|
"code": "scripts/utilities/upgrade_pve8_to_pve9.sh # main upgrade orchestrator\nscripts/utilities/pve8to9_check.sh # pre-check (separate menu option)\nscripts/utilities/proxmox-upgrade-pve8-to-pve9-manual-guide.sh # 17-step runbook\nscripts/global/update-pve8.sh # PVE 8 worker (final patches)\n/etc/apt/sources.list # bookworm → trixie sed\n/etc/apt/sources.list.d/proxmox.sources # created (deb822 PVE 9)\n/etc/apt/sources.list.d/ceph.sources # created if Ceph\n/etc/apt/sources.list.d/*.list # legacy files commented out / removed\n/var/log/pve8-a-pve9-<timestamp>.log # full upgrade log"
|
|
|
|
|
},
|
|
|
|
|
"references": {
|
|
|
|
|
"heading": "Official references",
|
|
|
|
|
Manuel Escalante