Proxmox Backup Server (beta)
- Thread starter martin
- Start date
can Proxmox Backup server is able to backup Proxmox VE individually or along VMS on it ?? also what is best approch to update proxmox VE , i mean how to backup proxmox VE before updating it to new release , in case something bad happen during upgrade.
Hi,
Thx. Good to know! In my use cases, it is not a nonsense at all ! But I agree that for others, where is in place 100 % trust relation it will be OK!
Good luck/ Bafta
Last edited:
Looking at the roadmap, they intend to include this function in the future. (https://pbs.proxmox.com/wiki/index.php/Roadmap)
In my home lab, this command worked to back up the Proxmox Host
Code:
proxmox-backup-client backup proxmoxRoot.pxar:/ --include-dev /etc/pve --verbose --repository root@pam@192.168.X.X:DatastoreLocalTo restore I set up this script, I believe it can be much improved
Bash:
# Variaveis
snap="host/pveb2/2020-07-29T17:24:14Z"
usuariopbs="root@pam"
ippbs="192.168.X.X"
datastorepbs="DataStoreLocal"
# Parar serviços
echo "------------ Parando serviços ------------------------"
for i in pve-cluster pvedaemon vz qemu-server; do systemctl stop $i ; done
# Exibindo backups
echo "----------- Exibindo backups --------------"
proxmox-backup-client snapshots --repository $usuariopbs@$ippbs:$datastorepbs
echo "Enter para continuar..."
read
# Restaurando ETC
echo "----------------- Restaurando Etc --------------"
proxmox-backup-client restore $snap proxmoxEtc.pxar /restaurar/etc --repository $usuariopbs@$ippbs:$datastorepbs
echo ".. ETC Restaurado. Enter p/ continuar"
read
# Restaurando Lib
echo "----------------- Restaurando Lib -----------------"
proxmox-backup-client restore $snap proxmoxLib.pxar /restaurar/pve-cluster --repository $usuariopbs@$ippbs:$datastorepbs
echo ".. LIB Restaurado. Enter p/ continuar"
read
# Restaurando Root
echo "----------------- Restaurando Root -----------------"
proxmox-backup-client restore $snap proxmoxRoot.pxar /restaurar/root --repository $usuariopbs@$ippbs:$datastorepbs
echo ".. ROOT Restaurado. Enter p/ continuar"
read
# Copiando arquivos
echo "----------------- Copiando arquivos -----------------"
cp -avr /restaurar/etc /
echo "ETC copiado. Enter p/ continuar"
read
cp -avr /restaurar/root /
echo "ROOT copiado. Enter p/ continuar"
read
cp -avr /restaurar/pve-cluster /var/lib/
echo "pve-cluster copiado. Enter p/ continuar"
read
echo "------------ Iniciando serviços ------------------------"
for i in pve-cluster pvedaemon vz qemu-server; do systemctl start $i ; done
echo "Enter para continuar"
read
# Iniciar o cluster
echo "Fazendo alteração para iniciar o cluster não vazio"
mv /etc/pve /etc/pve.bak
systemctl restart pve-cluster
echo "Finalizado.... Enter para fechar"
read
echo "----------------- FIM -----------------"in my tests, it worked
you can see the changelogs on the update GUI, if you already installed packages via 'apt changelog <packagename> or you can see changelogs via http://download.proxmox.com/debian/pbs/dists/buster/pbstest/binary-amd64/
not recommended (but possible). Best practice: local ZFS storage
see also https://pbs.proxmox.com/docs/administration-guide.html#backup-server-management
Imaginate, we have about 10 single PVE hosts with +300vms/cts, and remote internet hosts. In each PVE we have backup job shedule for vm and ct. In remote hosts we have cron jobs for backup (like db backup or user files, not all host). Its extremely hard to syncronyze all this backups, so that not owerload network or disk system. Is it possible to initialize backup job from PBS server, not from client? The idea is to create unified backup schedule for pve hosts and remote clients. Like as PVE backup processes, next host dont starts backup until previos ends. Can we expect this in PBS?
No. The backup server has no access to clients or knowledge about where the are - that's a feature and won't change.
Allowing one to configure that after a certain amount of currently active, incoming, jobs new clients need to wait until another finishes can be done without giving the server access to the client.
Having a schedule over multiple clients could also just get polled from the clients, so also if such a thing would be added we do not need to give the backup server full blown read access to all clients.
Both features aren't yet on any roadmap, but we will certainly consider them.
Nice! Love to see this long awaited feature!
btw. rust a proxmox, great!
btw. rust a proxmox, great!
The most honest answer is:
We'll working very actively on ironing remaining issues out of the base functionality and integration.
Other hints: https://forum.proxmox.com/threads/subscription-type-cores-for-vm.73793/#post-329605
vzdump of turnkey linux templates would not restore properly in unpriviledged containers because of postfix directory. There was a not so simple fix of excluding directories via CLI backup/restore. Is this process made simpler with the backup server?
https://forum.proxmox.com/threads/unprivileged-containers.26148/page-2#post-207631
https://forum.proxmox.com/threads/unprivileged-containers.26148/page-2#post-248550
HI,
For me Tuxis PBS crashed/was not available since yesterday 21:00 GMT +2
Again I had to restart pvestatd. I also had to kill all '/usr/bin/proxmox-backup-client status ' processes on one node
Glad to see that Tuxis PBS was upgraded to v0.9 (what's new for 0.9 ?)
For me Tuxis PBS crashed/was not available since yesterday 21:00 GMT +2
Again I had to restart pvestatd. I also had to kill all '/usr/bin/proxmox-backup-client status ' processes on one node
Glad to see that Tuxis PBS was upgraded to v0.9 (what's new for 0.9 ?)
Changelog is always a good read for that, depending on if you're on the client or server side use:
Bash:
apt changelog proxmox-backup-client
apt changelog proxmox-backup-serverThe Proxmox VE and also Proxmox Backup Server webinterface has a changelog button for pending updates in the webinterface.
If all fails, there's always: http://download.proxmox.com/debian/...amd64/proxmox-backup-server_0.9.0-1.changelog
Hope that helps.