ZFS: Automatic Snapshot including RAM?

[JohnD]

Hello everybody,

is there any way to configure an automatic snapshot inlcuding RAM ?

Thanks.

 

[JohnD]

I tried it with eve4pve autosnap and it works fine but it makes Snapshots without RAM.
Any ideas how to change that?

 

[Rhinox]

AFAIK, on PVE there is no way of creating snapshot of running VM (be it KVM or LXC) including VM-memory. Because of this, I still have to use vSphere/ESXi where VM-downtime is unacceptable, yet I must be sure my backups are consistent...

 

[tom]

AFAIK, on PVE there is no way of creating snapshot of running VM (be it KVM or LXC) including VM-memory

See https://pve.proxmox.com/wiki/Live_Snapshots

 

[JohnD]

AFAIK, on PVE there is no way of creating snapshot of running VM (be it KVM or LXC) including VM-memory. Because of this, I still have to use vSphere/ESXi where VM-downtime is unacceptable, yet I must be sure my backups are consistent...

The live snapshots are working absolutly great. Never had an inconsistent system after a restore.

I just do not understand why there is no way to do snapshots automatically. Is that at least on the roadmap? seems like a very basic feature for me.

 

[LnxBil]

I just do not understand why there is no way to do snapshots automatically. Is that at least on the roadmap? seems like a very basic feature for me.

Write a script that does it by calling qm snapshot for all your desired VMs.

For simple machines like Webservers (without DB), Desktops and such, VM snapshotting is great, but in an distributed production environment, this is a no-go. I'd never use VM-based snapshotting for HA. Best is always to run a service with its built-in ha capabilities e.g. for database clusters and such (e.g. Oracle RAC), synchronous replication is a must for zero-data-loss.

 

[fireon]

Backupsnapshotting with qemu-guest-agent should be not problem with DB's too. Because the agent is responsible that the backup is consistent. In Linux and also in Windows with shadowcopy.

 

[Taylor Murphy]

My system uses ZFS and i have configured the package "zfs-auto-snapshot" to take a snapshot every 5 minutes of every layer of my zpool and then uses intelligent storage to keep 1 copy of each year, 2 monthly, 4 weekly, 7 daily, and 24 5 minute backups. I also have another node running ZFS that I set up a replication schedule in the GUI to replicate my most important VMs every 5 minutes and the rest of them every 2 hours or daily. I haven't had any issues yet (except that I can't figure out how to set up a replication schedule for one of my LXC containers that has a bind mount passthrough).

 

[fireon]

My system uses ZFS and i have configured the package "zfs-auto-snapshot" to take a snapshot every 5 minutes of every layer of my zpool and then uses intelligent storage to keep 1 copy of each year, 2 monthly, 4 weekly, 7 daily, and 24 5 minute backups. I also have another node running ZFS that I set up a replication schedule in the GUI to replicate my most important VMs every 5 minutes and the rest of them every 2 hours or daily. I haven't had any issues yet (except that I can't figure out how to set up a replication schedule for one of my LXC containers that has a bind mount passthrough).

Really, really nice!

 

[davidpellerin]

My system uses ZFS and i have configured the package "zfs-auto-snapshot" to take a snapshot every 5 minutes of every layer of my zpool and then uses intelligent storage to keep 1 copy of each year, 2 monthly, 4 weekly, 7 daily, and 24 5 minute backups. I also have another node running ZFS that I set up a replication schedule in the GUI to replicate my most important VMs every 5 minutes and the rest of them every 2 hours or daily. I haven't had any issues yet (except that I can't figure out how to set up a replication schedule for one of my LXC containers that has a bind mount passthrough).

Love it. Thank you so much for this info. Good bedtime reading ahead!