Discussions related to using object storage as a backup target.
Post Reply
geets
Enthusiast
Posts: 25
Liked: 2 times
Joined: Mar 03, 2020 2:50 pm
Contact:

Changing backup plan completely

Post by geets »

Hello
I am planning a change at a Backup site and would be grateful for some input.
First I'll outline how the current setup is and why we want to change things so you get a broad overview:

Currently there are 13 VMs on a single ESX-Host. They get backed up by veeam agents. (Thats a temporary solution).
Those VMs consume about 8.5 TB on the ESX-Host

They get backed up daily by the agent to a QNAP-NAS with a size of about 15TB
On the backup target they need about 8TB of space - a little less than the disk files on the ESX-host but I heard that could happen due to veeam magic.
So theres 7TB free space available on the backup target.

Right now the backups are reverse incremental. That means there is one full backup file (.vbk) for every VM and 14 incremental backup files (.vib) on the NAS.
There are no Copy Jobs configured.

I hope the above outline helps to paint a mental picture of the system.
I'll get to the changes we are planning to do now and hope you share your thoughts on them:
  • There will be cloud-storage involved to off-load long term retention backups. (that aren't present as of today)
  • In order to do this we configure an SOBR and offload to the cloud (S3 compatible bucket) everything thats older than say 30 days.
  • The local extend of the SOBR will be on the current NAS.
  • We plan to have a daily backup job that creates 31 restore points on the local extend of the SOBR.
  • Every saturday a synthetic full backup will be created. This is necessary so the offloading process to the cloud can take place (it needs a finalized backup-chain if I understand correctly)

- Please share if you think this is a reasonable idea. I'm doing this for the first time.

Also I have concerns about the storage space we need on our NAS if we configure it this way but the final consumption is too hard to predict for me.
Right now our backups are of the reverse incremental strategy. That means we have one full backup and 15 incrementals.
This will be replaced by four full backups and 26 (or so?) incrementals in between the full backups.

- Will we really need about four times the storage space when we change our backup strategy as I described?
What do you think?
PetrM
Veeam Software
Posts: 3208
Liked: 515 times
Joined: Aug 28, 2013 8:23 am
Full Name: Petr Makarov
Location: Prague, Czech Republic
Contact:

Re: Changing backup plan completely

Post by PetrM »

Hello,

I think it's a definitely good idea to add an object storage and start using Capacity Tier as an additional tier of storage especially if you're running out of space.

But I would opt for increasing a gap between retention period (31 restore points) and operational restore window to gain some storage space for performance extents.
For example, you could offload backups older than 14 days, just make sure that you have enough space for 2 .vbk and 12 .vib as long as you run backups daily and fulls are performed once per week.

You may take a look at restore point simulator to estimate an amount of free space required on performance extents.

Thanks!
geets
Enthusiast
Posts: 25
Liked: 2 times
Joined: Mar 03, 2020 2:50 pm
Contact:

Re: Changing backup plan completely

Post by geets »

Hi Petr,

thanks for your input. I think we will have to do as you suggested and save space on the local storage by offloading earlier.
Here is another point of concern:
I just realized there is one single machine that consumes 7TB of space. An initial upload into the cloud will take days.
Do you see any issues from a veeam/technical standpoint? E.g is there a chance, veeam might abrupt the upload if it takes longer than a given time?
What if it would take longer than a week?

Thanks in advance
veremin
Product Manager
Posts: 20258
Liked: 2248 times
Joined: Oct 26, 2012 3:28 pm
Full Name: Vladimir Eremin
Contact:

Re: Changing backup plan completely

Post by veremin »

Nope, there are no issues with the described scenario - Capacity Tier is designed to sustain long offload periods and periodic network issues. Thanks!
geets
Enthusiast
Posts: 25
Liked: 2 times
Joined: Mar 03, 2020 2:50 pm
Contact:

Re: Changing backup plan completely

Post by geets »

PetrM wrote: Apr 30, 2020 3:30 pm Hello,

I think it's a definitely good idea to add an object storage and start using Capacity Tier as an additional tier of storage especially if you're running out of space.

But I would opt for increasing a gap between retention period (31 restore points) and operational restore window to gain some storage space for performance extents.
For example, you could offload backups older than 14 days, just make sure that you have enough space for 2 .vbk and 12 .vib as long as you run backups daily and fulls are performed once per week.

You may take a look at restore point simulator to estimate an amount of free space required on performance extents.

Thanks!
Would it be possible to save even more space by offloading backups older than 7 days?
Or do we need 2 .vbk in any case since backup chains only qualify for offloading when there is another .vbk in place
dalbertson
Veeam Software
Posts: 492
Liked: 175 times
Joined: Jul 21, 2015 12:38 pm
Full Name: Dustin Albertson
Contact:

Re: Changing backup plan completely

Post by dalbertson »

You are correct in stating that you would need a new vbk before the older backups would be offloaded....but you could still save space as soon as the new vbk is created the next time the offload process happens it will tier the older ones off. So there would only be a short period of time where there would actually be 2 vbks
Dustin Albertson | Director of Product Management - Cloud & Applications | Veeam Product Management, Alliances
backupquestions
Expert
Posts: 186
Liked: 21 times
Joined: Mar 13, 2019 2:30 pm
Full Name: Alabaster McJenkins
Contact:

Re: Changing backup plan completely

Post by backupquestions » 1 person likes this post

Another thing is, if you are using refs or XFS with reflink enabled as your local nas repo file system then veeam can use block cloning so that despite having multiple synthetic fulls, they are spaceless and only consume the size of an incremental. This way there is no real issue with having 2 or even 3 fulls on disk while you wait for offloads.
Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests