Space efficient long term backups w/o tape

[loelly]

So, I'm currently wondering what your preferred way of keeping long term backups would be. We need a year of retention, but not daily granularity for some bigger file server virtual machines (8 TB net).

Currently, with our old and outdated Veeam B&R 8.0 infrastructure we're keeping a reverse incremental backup chain of around 180 days and switch storage devices thereafter. This was done to cope with limits on the storage side and yes, we are aware of how bad such a long (reverse) chain of backups is.

Now that we're moving to Veeam 9.5, I get a chance to completely redo the setup, but am also limited on the storage side, but can do ReFS (Windows Server 2016, ReFS repo on affordable but not do bad SAS backend - 24x4 TB HDD, dual controller, quite some cache). We've switched all jobs to Forever Forward Incremental with weekly synthetic fulls on ReFS. Now my idea was either to use a backup copy job to a new location (on same storage), and set up weekly and monthly retentions. As this makes use of ReFS, it should not consume a lot of storage (1x initial copy + a weeks delta plus ReFS pointer overhead, right?).

On the other hand, I am wondering if there's any downside in having a long Forever Forward Incremental chain with weekly Synthetic Fulls? As the change rate is not so much compared to the net capacity (50 GB/day), we could just go with this setup...

Any thoughts?

[foggy]

Hi Jens, depends on the length of the chain, but I'd definitely use backup copy for longer retention and introduce some sort of second storage to your setup, since having all the backups on a single storage is not recommended.

[Gostev]

For a long term archive, you do want to use Storage Spaces mirror or parity set - if ability to recover from archive is important

The need for second storage is kind of perpendicular to retention: you need this regardless of whether your retention requirements are short or long term. Normally, you'd have a primary backup repository (fast storage, short retention to cover operational recoveries) - and a secondary backup repository holding backup copies (cheaper slow storage, long retention to cover your data retention requirements).