How To Best Handle Backups

[tegroup]

I have 27 vms and physical servers I need to backup daily.

Every day during the month, 1 of the vms or physical servers does a full backup, the rest do incremental.

Some physical servers are far away or have lots of data.

We're using a QoreStor Deduper Hardware Appliance.

We were using the QoreStor as Object (S3) Storage, but we found the cleaner couldn't keep up with the data.

We've moved to using the QoreStor as EDM (Enhanced Data Mover), but found we can no longer run multiple backups at the same time.

Another problem with EDM is that if a backup takes too long, and it becomes 00:00, the next day, the full backup from yesterday won't trigger, as it's now the next day.

What should I do to scheudle all of our backups to ensure we're getting a full backup every 30 days per vm?

[rennerstefan]

Hi

its very hard to provide some useful guidance here without more details on your environment.
You are saying "Some physical servers are far away", does that mean you run the backups from a central place via a WAN connection?
If so what does "lots of data" mean in terms of GB/TB and what is the bandwidth?

Would be great if you could provide some more details on those 27 machines and the design you have.

Thanks

[david.domask]

Hi tegroup,

My instinct is to tell use Synthetic Fulls, but reading the QoreStore documentation looks like it doesn't support fast clone; fast clone is not essential for Synthetic Fulls, but it helps immensely with backup times as noted in the User Guide page.

I would still give it a shot as the main benefit is you only need Incremental backups (helps with your distant sites), but it will heavily depend on the performance of the QoreStore repository and if it can handle the workload.

Out of curiosity, what's the bottleneck shown as in the jobs during the long running Active Fulls?

[haslund]

How many backup jobs do you have? Are you scheduling all full backups on the same day?

[tegroup]

We're currently using 27 backup jobs to backup the 27 various VMs/physical servers.
--- Can someone suggest a better way to backup our things?

We just switched our Hardware Deduper from S3 to Enhanced Data Mover.

We thought EDM would only let us run one backup at a time, we're unsure currently.

Q: You are saying "Some physical servers are far away", does that mean you run the backups from a central place via a WAN connection?
Our Veeam Server and Qorestor Deduper are local to city x. Some servers are across the country in city y. Some are in Azure.
about half of the servers are local to Qorestor/Veeam. Everything is connected with Bell MPLS. Azure servers are part of the company wide WAN(Bell MPLS) too.


It would be awesome if I could schedule the jobs so the 1 full backup (changes every day) ran first, and then all the sequential triggered, but I don't know if we can do that.

[haslund]

You could start by considering whether all your VMs really have different data protection needs. For example, are the RPOs different for all 27 VMs?
If the VMs are in different locations, you could consider creating a single job per location for the VMs.

For the full backups, consider scheduling them for different days so there is no overlap.

[tegroup]

Yeah, each day a different VM is being backed up with a full backup.
First Monday VM1
First Tuesday VM2
...
Fourth Friday physical server 6 (27)


We have 1 physical server in city 1, we have another physical server in city 2, ... another physical server in city 5.

I'm not seeing how using a single job would help, I apologize. I'm not familiar with RPOs.