Partitioning for ProxmoxVe 4

[Ovidiu]

I'm playing with a new server I'd like to run Proxmox 4 on - hosted by SoyouStart which offer an automated installation of Proxmox 4.
The default partitioning seems to lead to this:

Device        Start        End    Sectors    Size Type/dev/sda1        40       2048       2009 1004.5K BIOS boot/dev/sda2      4096   40962047   40957952   19.5G Linux filesystem/dev/sda3  40962048   43057151    2095104   1023M Linux swap/dev/sda4  43057152 3905935359 3862878208    1.8T Linux LVM

its a 2x2TB HW Raid. Any problems with this or should it be changed? Will I be able to backup like this or do I need separate storage for backups?

 

[t.lamprecht]

I would say the space is enough for backups although it's depending on how much and how often you want to backup.
You have to note that making backups on the same disk/storage where the Virtual Machines and Container disks reside is not a good practice.

If the storage fails everything is lost, the backups won't help.

# df -h

outputs a more detailed view of what space is free, the pve-data LV is the important one for backups and VM disks.

 

[Ovidiu]

Thanks for the reply and sorry I should have been more specific.
I know the best practices for backups, the reason for my question was that reading through the forums I found different mentions that different types of backups suspend/stop/snapshot lvm backups need different types of storage and some won't work with certain configurations.

So what I'm asking is:
planning to have a few LXCs and possibly 1 windows-based KVM further down the road, does this partitioning work? I'm aiming at doing backups with no down-time.

I do have external storage for backups, i.e. FTP and I also use duply/duplicity to backup externally.

Does this extra info help clarify my question?

df -h
Filesystem            Size  Used Avail Use% Mounted on
udev                   10M     0   10M   0% /dev
tmpfs                 6.3G   17M  6.3G   1% /run
/dev/sda2              20G  2.3G   16G  13% /
tmpfs                  16G   40M   16G   1% /dev/shm
tmpfs                 5.0M     0  5.0M   0% /run/lock
tmpfs                  16G     0   16G   0% /sys/fs/cgroup
/dev/mapper/pve-data  1.8T  4.1G  1.7T   1% /var/lib/vz
/dev/fuse              30M   12K   30M   1% /etc/pve
cgmfs                 100K     0  100K   0% /run/cgmanager/fs

 

[t.lamprecht]

KVM/QEMU live backup should work on all storage types, a reliable (and not the slowest) connection helps to avoids issues.

Making a Snapshot works also on all storage types if the qcow2 disk image format is used.
Else, when using raw, the storage needs snapshot support (zfs, ceph to name a few).

So you're "problem" will be the LXC Containers. If you want completely life Container backups you need an storage which supports snapshot (copy on write), ZFS (you can setup PVE on ZFS) could be an option for you.

Else suspend mode backups work on every Storage/FS and it will minimize the downtime, but you will have one.

 

[Ovidiu]

Thanks for taking hte time to help out. I have also read all the relevant pages in your wiki but was still not 100% hence this thread.

Ok so basically as is: 2x2TB HW Raid and Proxmox installed on there there is no storage supporting snapshotting an LXC.
What if I shrink the space currently occupied by Proxmox, could that space be somehow used to do snapshots of LXC?
I mean ZFS can't be used on top of HW Raid, any other possibility that works out of the box with Proxmox and my current hardware?
I guess I could live with suspend, but first I would like to get confirmation that the way I have partitioned my resources is the optimal situation before I go live and start actually working on this server.

Slightly embarrassed but I just realized the that the "snapshot compatible storage" refers to the storage where I store the VMs not where I am backing up to, right?