Inherited a Nightmare and need help making it a dream

[NickSCA]

Working on a system for a medium sized Medical clinic. Current backups are done with Reverse incrementals and the Backup Storage drive has over 80% fragmentation and can't be defragged due to the constant load on the disks.

Current Environment:

• 2 Hyper V Hosts
  20 VMs
  Regular Veeam Backup and Replication (Not Enterprise version)
  Veeam Backup connected to a MD1400 DAS

I keep on getting inconsistent backup errors that are all over the joint and other fun things with the backups and I'm thinking of just doing a nuke and pave on the backup server and starting from scratch.

1.) If I do a Nuke and pave (I have good offsite backups), will ReFS give me a better chance against the fragmentation on the Backup Storage Drive
2.) Should I go with the recommended path (I have about 30TB of space on the backup drives and right now I'm sitting at 17TB of space used with 12TB left) of incremental with Synthetic Full backups and an active full backup quarterly?
3.) I'm hearing mixed things on whether or not to do the Volume Snapshots. You guys have any opinions on it?
4.) Guest quiescensce... Yes or no?
5.) Secondary Destinations or Backup Copy?

I have my opinions on this, but before I head down this road, I would like to get a few brains that are smarter than I.

My opinions are as follows:
1.) Yes to ReFS due to the Salvage mode and the way it handles the Metadata (Doesn't update in-place, creates copy and once it's happy with the results does it update the metadata). This turns what is usually a server that is brought offline for time consuming disk checking utilities to find and repair the entries, to a volume which is repaired except for the corrupt data files and brought back online in under one second.
2.) Incremental with Synthetic Fulls and Active Fulls once a quarter
3.) On the fence, but I do say yes to the snapshots.
4.) Yes with the exception of SQL and Exchange, there you turn on the Application aware processing.
5.) Either secondary location or backup disk, don't have the disk space for both I don't think.

Limitations:
The environment is solid, I can't add new disk hardware at this time. (1.8 million on a surgical center takes priority)

To all, thanks in advance for your help and opinions! And all of you Veeam Software Gurus (Foggy and Gostev Especially) I'd love to hear from you!

[MichaelCade]

Hi Nick,

Can i ask what your production storage is also? but i am sure i can offer some guidance for some of the other parts.

ReFS is pretty topical at the moment and lots of people are looking at using the file system especially for Veeam backup files, the caveat is though its new from Microsoft and there have already been a few bugs found.

Back to ReFs, its for sure going to help you when it comes to the merge and synthetic operations as it uses pointers rather than taking a complete copy of the blocks for the synthetic full. check this out for a visual - https://www.veeam.com/blog/advanced-ref ... suite.html but as i said just the word of warning is around its maturity as well as you mention fragmentation, ReFS is still prone to framentation but only with large block sizes.

The volume snapshots point i am going to park till i know what production array you have.

Application Guest quiescensce - if you are looking for transactionally consistent backups or replicas then yes you should enable. also if enabled you could leverage the "All editions provide a catalog of guest files in backups currently on disk."

Secondary destinations, i presume the above is more of your local backup target looking to hold that 7/14/30 day retention and then you are also looking for an offsite location to store these backups but also possibly more? really we dont mind what you use, just advise to adhere to that 3-2-1 ruling. But we do also have the option for you to send those backup copy jobs to a Veeam Cloud Service Provider offering Veeam Cloud Connect Backup, this negates the cost of a secondary disk array and leverages an OPEX model cost instead with the service provider. We also have the ability to go to a tape device to provide you with an air gapped copy of your backup files on a different media type and offsite.

You mention the "cant add new disk hardware" so possibly the cloud connect operational model may fit into the budget.

Hope that helps a little.

[NickSCA]

Michael, Production storage is basically attached directly to the servers. No NAS, DAS, or any external Raid array

Basically I have 2 Dell R 730s with 1.2 TB 15K's on them. I believe it's a Raid 10 array on a PERC 700 series if that helps, I'd have to recheck to make sure.

And Thanks for the article... I really appreciate that eye opener!!

[NickSCA]

Oh and the CEO of the Medical Clinic is a "If I can't see it, I can't control it" kinda person so Cloud is out of the Question unfortunately!

[MichaelCade]

OK understand. When you said snapshots were you referring to native hyper v snapshots if so I don't think these are going to offer you the level of recovery you can achieve with Veeam.

[NickSCA]

I was referring to the snapshots within Veeam. I forgot that we don't have telepathic links to forums like these, and I thought that everybody can read my mind because in my mind it made sense. I think I'm gonna have to work on the telepathic stuff after I fix these backups!

[MichaelCade]

So in your environment we would leverage the hyper v snapshots to take a backup to disk, we can then either take a backup copy or Tape Job for an off-site copy and possible GFS restore points. We don't have our own snapshot functionality.

[foggy]

Hi Nick, if by volume snapshots you meant the ability to backup from storage snapshots, then this Veeam B&R feature is only available for VMware. For Hyper-V backup, Veeam B&R uses technically similar functionality that is provided natively by Hyper-V - off-host backup from Hyper-V volume snapshots with the help of hardware VSS providers (Veeam itself hasn't implemented any "integration" in this regard).