Veeam Copy Jobs - Incrementals

[collinp]

Our physical servers are backed up using a different backup software with the following retention:
12 Monthly Fulls
48 Weekly Differentials
90 Days of Incrementals (to match Glacier)

We also have a very large Veeam backup environment to backup our virtual machines. How would we match the above physical server retention using Veeam Backup Copy jobs to our Cloud CIFs appliance?

Option 1: This is going to be costly in the cloud due to the number and size of Weekly backups:
52 Weekly Backups
90 Days of Incrementals

Option 2: I would prefer this, but I don't see a "long-term" archiving incremental option with Veeam Backup Copy jobs. Even though the coverage is better than the above Veeam Copy design, the size would be much smaller:
12 Monthly Backups
365 Incremental Backups


Thank you!

Collin

P.S. Our primary Veeam Backup jobs retention is 1 Monthly Active Full and 30 daily incrementals. Our data volume shrunk enormously when we changed to this method over storing weekly fulls

[collinp]

I thought of a better way to explain this. Essentially what I want is a Simple Retention Policy in my Backup Copy Jobs containing 30 Restore Points. However, I don't want Veeam to delete these Restore Points after the 30 days. Instead, after 30 days I want a new Monthly Backup chain created. I want to store 12 Monthly chains consisting of 1 Full and 30 incrementals in each. This way, over a years time, I would only be storing 15 copies of my data in Glacier. Versus if I had to do 52 weekly full' and storing 52 copies in Glacier - ouch.

I need the Backup Copy option to do this because within the Copy jobs I have the option to have dedup and compression set to none since our Cloud CIFs appliance dedup's across all jobs.

Collin

[veremin]

Wouldn't it better in this case to create additional backup job with 365 increments and monthly full backups?

Though, I have a couple of questions:

- Why do you need to store such a big number of increments on storage that resembles mostly tapes (archiving storage)?
- What particular device you're talking about?
- I'm wondering because Glacier itself is more like "freeze and forget your data", so, how does the device guarantee access to previous incremental files required to create subsequent one?
- How does it guarantee fast access to them, if restoration need arises?
- Aren't you concerned with the price Amazon is going to charge you with, should you require to download data from Glacier? Or the price is included anyhow in the said device offering?

Thank you in advance for the feedback.

[collinp]

We are using AltaVault appliances (Steelstore). We have a 10Gbps pipe to AWS. 30 days of incrementals are stored in S3. Through a Lifecycle policy the backups graduate from S3 to Glacier after 30 days. In our experience with using these appliances, we have had no issues with restore performance from S3 or Glacier. I can even mount S3 and restore individual files from the backups from Veeam. Backup performance in our testing was over 980 Mb/s. There is a local cache where data is stored on the appliances as well.

Our requirement is to keep 30 days of the most recent backups and 52 weekly backups. We don't have to keep 365 incrementals but I have found that this is the most space efficient way to store a years worth of Veeam backups offsite based on our experience with tape and based on our requirements. Our monthly change rate is 20%.

Example 1 (Stores 62 copies offsite. 1TB of data becomes 62TB at the end of the year)
-52 Weekly
-30 Daily

Example 2 (Stores 15 copies offsite. 1TB of data becomes 15TB at the end of the year)
-12 Monthly
-365 Daily

I should mention that Veeam Backup Copy jobs are not a requirement. The Altavault could act as the primary backup target for Veeam as well. Either method would work: Altavault for Veeam Primary Backup job or Altavault for Veeam Copy job. I just need a method that won't delete my incrementals. I was thinking of doing a Backup Copy Job with Simple Retention and the rotate drives option checked on the reppository. At the end of the month, rename the full incrementals so the Copy job won't delete them. Then the Copy Job will just produce a new backup chain. Networker/Backup Exec/Commvault have the concept of differentials so that weekly backups could be stored efficiently without having to produce a full every week. We got around this with Veeam by using the Example 2 method above which has been very space efficient for us on tape.

Collin

[collinp]

If we were to use regular Backup Jobs (not a copy job) configured to run a Monthly Active Full and 30 incrementals, writing straight to the AltaVault CIFs share, is there a way where we can configure the job not to delete any of the backup chains.

The issue is that the Altavault CIFs share maps directly to S3. If the job deletes a chain from the CIFs share, it deletes it from S3.

I want the CIFs share to look like this:

January Active Full
January 30 incrementals
February Active Full
February 30 incrementals
March Active Full
March 30 incrementals
etc....for 1 year. A script can kick off daily to watch that files get deleted after 1 year of retention.

Collin

[veremin]

With retention equal to 365 a backup copy job won't start to purge existing chain during a year. After one year, it will start to delete those restore points that are already falling our of the specified retention.

So, additional backup job with 365 increments and monthly full backups sounds like the best idea.

Thanks.