Your root file system doesnt really matter for the purposes of this discussion. only the vm storage does. Assuming you intend to use the same filesystem for your OS and payload, you cant use ZFS replication- but that doesnt mean you cant...
No, since it needs a filesystem-specific feature which ext4 or XFS simply doesn't have. In theory ut would je possible with btrfs but at the moment ProxmoxVE storage replication only works with ZFS.
I would backup all vms/lxcs on your node (...
In hindsight, you should have skipped the HBA and SAS and gone with NVMe.
When choosing between ZFS and Ceph, keep in mind that ZFS is a local filesystem, while Ceph is a form of distributed/shared storage. Each comes with its own set of pros...
Enteprise Grade SAS drives -> SSDs?
If you decide to go the ceph route you should probably upgrade your network to 25G or faster if you have fast SSD's, otherwise the network could be a performance bottleneck.
Here are some official Ceph...
So you’ll have 5 servers? You can use Ceph with that . It would be ideal to use one NIC for Ceph public and one for Ceph private. Corosync should use both.
Will you be using a server for PBS?
The new ZFS version (that comes with the new Proxmox) has a default maximum of 90% of memory, while the previous versions had a default maximum of 50%. If you never set a memory limit before, this is a big change. Maybe manually configure a limit...
Hi,
LVM-thin ist da speziell und Klone von Volumes sind automatisch eigenständig, i.e. sie referenzieren die gleichen "Extents" wie das originale Volume, aber sind nicht davon abhängig dass das original weiter existiert. Wenn dann das originale...
Hey,
Wireguard[1] or other VPNs is probably what you are looking for. Depending on whether you have a static IP you may also need DDNS or some small VPS somewhere with a static IP that your VPN clients all connect to.
[1] https://www.wireguard.com/
I believe this is not implemented in the Gui (yet?).
On commandline it works as intended. man zpool-attach tells you the syntax.
This is a manipulation on the lowest level and failures may be fatal. Personally I do set a global checkpoint to...
For future reference...
https://github.com/kneutron/ansitest/blob/master/proxmox/proxmox-storage-report.sh
Lots of other helpful scripts in that repo. I would run the above nightly in cron; and also strongly recommend running the bkpcrit script...
Buy some USB-to-SATA adapters to use them temporarily for this task. They are cheap, five to ten Dollars/Euros each...
Edit: that is probably not going to work for 3.5 inch harddisks, because of the required power...
You could install needrestart with apt install needrestart. Then run needrestart after an update and you will get notified whether a reboot due to a new kernel is necessary
PDM enters beta - it may possibly help migrating between non-clustered systems. (I did not test PDM yet...)
https://pve.proxmox.com/wiki/Category:Proxmox_Datacenter_Manager
Announcement 30 minutes ago...
Including the ProxmoxVE configuration or just the lxcs + vms? For the VMs and LXCs you have pve-zsync and the remote-migrate commands of the pct and qm cli tools:
https://pve.proxmox.com/wiki/PVE-zsync
https://pve.proxmox.com/pve-docs/pct.1.html...
Wenn Geld da ist, dann 4x DC600M Serie Enterprise SATA 3.0 SSD, alle vier SSD bilden einen ZFS Pool, den man als 2 VDEV ZFS mirror erstellt.
VDEV0 ZFS mirror 2x ssd - Stripe - VDEV1 ZFS mirror 2x ssd
Dadurch nutzt man maximal, ca. 80 %, der ca...
Hallo,
was mir nicht ganz klar ist: Wieviele compute-Server setzt ihr ein? Da ihr auch eigene Storage-Hardware habt, gehe ich davon aus, dass ihr einen Cluster habt, je nach dessen Größe muss man schauen.
Zum Rest: Ich habe nicht wirklich...
I really tried to find an existing thread, but everything I find is either related to the main/boot pool, or adding an encrypted dataset only, so I figured I write down my steps for my own documentation and maybe it'll help someone else out too...
That's of course the problem here. When I steal that computer I get the key for free.
What attacking vector is it you are securing with this approach?
The only situation I see is if the devices die so hard, that the build-in "erase"-function...