Coherence update 2022
|Coherence update 2022|
|Name||Coherence update 2022|
- Date still needs to be determined.
We are several versions behind on Proxmox on Coherence, as of writing the current version of Proxmox is 7.1-7 while we are still running 5.4-13(!). This means that we can't make any new LXC containers with newer distributions because they are simply not supported by Proxmox 5. Several attempts to upgrade Coherence have been made in the past, but I hope this time we can actually do it! It would also be nice if we could use this as an opportunity to clear old unused VMs and containers off Coherence.
Upgrading Proxmox path
Upgrade stuff (can be done already when backups are possible)
Please make backups first before you upgrade a VM or container
There should be a backup system available on backup-nfs before we start.
Debian containers should be Debian 10 at least before we can upgrade Proxmox, and Ubuntu 18.04 it seems.
Thinking steps required
* Backups - if we mess up, how do we bring things back?
- Need at least 4Tb to do this.
- Redundancy - Can we upgrade some other host and migrate? Would this be a good plan going forwards?
- Probably not required - Proxmox upgrade scripts should check containers before upgrades.
- Upgrades - All VkMs/Containers below need to be ticked off
A system called `notnotapp` has been setup and has 4x 1TB in raid 5 storage, meaning we have 3TB available for backup with redundancy. Although, roughly 1.8TB has been used for the mp3 share and the user data. Which should leave enough for VMs however, it is advised to remove your backup once the VM has been successfully upgraded to leave enough room.
A update of the user directory and mp3 share has finished at 10 January, rsync should be ran from time to time to keep notnotapp in sync with notapp. There is 966GB left free for VM backups.
List of LXCs/VMs
LXCs/VMs in need of OS upgrade
|100||LXC||rollout.vm.nurd.space||Debian 9||-||Petraea - a junky attempt at trying to come up with a common deployment standard for Nurdspace, and some mechanism of applying security standards. Could be folded.|
|101||LXC||notapp.vm.nurd.space||Debian 9||-||Critical - this is the main container that provides storage mounts to everywhere (egg, mpd). Update this one early.|
|102||LXC||parts.vm.nurd.space||Debian 9||-||A partkeepr frontend. Zarya?|
|103||LXC||postgresql.vm.nurd.space||Debian 9||-||Is seriously broken :< needs fixing (plz save the db)|
|140||LXC||nurdbot.vm.nurd.space||Ubuntu 16.04 LTS||-||Crticial|
LXCs/VMs that are upgraded
Some of these have been done previously
|105||VM||spacesound.vm.nurd.space||Ubuntu 20.04.3 LTS||OK|
|107||LXC||nurdbotdev.vm.nurd.space||Ubuntu 20.04 LTS||OK|
|123||VM||jarvis.vm.nurd.space||Hass OS 6.6||-|
|124||LXC||sdr.nurd.space||Ubuntu 20.04 LTS||OK|
LXCs/VMs that are not running
These can possibly be deleted or archived.
|110||LXC||bridger.lan.nurd.space||?||-||Petraea - An old routing box for prototyping different network layers. Could be folded.|
|113||LXC||panopticon.vm.nurd.space||?||-||Petraea - An old monitoring setup (zabbix). Could be folded.|
|114||LXC||vinculum.vm.nurd.space||Debian 8||-||Petraea - An old IoT processing system. Could be folded.|
|115||LXC||wopr.vm.nurd.space||?||-||Gaming VM. Zarya?|
|137||LXC||wol-server.vm.nurd.space||Debian 7||-||Petraea - Some central means of coordinating WoL handling with an API. Could be folded.|
|143||LXC||neural.vm.nurd.space||?||-||Petraea - an old artistic neural network processor. Could be folded.|