Calculate Restore Points correctly

[dsimpkins]

Hi all,
have been spending some time searching the boards for answers on a few things, though i'd simply ask the questions straight up instead:

1. If it's ok, I want clarification (and recommendation even) on setting the amount of restore points on a file server. I have a file server, i want a backup of changed files every 30 mins between 8am-6pm (10 hours). I also want to keep 30 days of restore points. My calculation is that 10 hours of half-hourly increments is 20 restore points per day, times by 30 days equals 600 restore points. My two specific questions here are: is this correct, and is this a reasonable way of going about backup?!

2. an extension of the above, if i use the "continuous" mode, how would I keep 30 days of restore points?

In both scenarios, after the first day (8am-6pm) of half-hourly increments, im happy to have one final backup and lose the half-hourly ones, so what i mean is, at 5:30pm on the last day of a month I would have 29 daily restore points plus the last day's 19 half-hourly restore points.

Gee, this is difficult to explain in a forum post! Im hoping someone can guide me on the best path... I recall our implementer talking about the "daily full backup" option somewhere here, maybe that plays a part.
Thanks
Daniel

[foggy]

Daniel, are you sure you really need that frequent backup of the file server? Typically file servers' content is pretty static and does not change much during one working day. Do you have a high change rate on this file server? I would suggest a daily backup of this VM with 30 restore points for a monthly retention. Thanks.

[dsimpkins]

Hi Alex, thanks for your reply. Yes the daily backup with 30 restore points is exactly what I have now.

Due to the nature of our data-intensive work, the rate of change is what I would consider quite high (perhaps others see this differently), eg, over the last 30 days the average daily reverse incremental backup is 6GB (there are peaks of 19gb and troughs of 1gb).

The 30 minute increments are pretty much to protect against accidental deletion, it's something that the users had prior to me joining the company, and something they were used to, so if possible I want to add this protection again.

So would my plan work? Better way of doing it?

[foggy]

Well, what you can do currently via Veeam B&R UI (if you need 20 restore points per day resulting in a daily full), is to set up forward incrementals with 1 restore point and everyday active full, schedule the job to run periodically every 30 minutes and open 2 backup windows for this job: 1am-2am (for example) and 8am-6pm. This will run active full at night (deleting previous incrementals) and incrementals through the working day. Additionally you will need to offload the full backup daily once it is created and maintain 30 days retention manually or via script. Alternatively, you can launch your jobs with a PowerShell script using Windows Task Scheduler and set up the required schedule.

[Vitaliy S.]

The 30 minute increments are pretty much to protect against accidental deletion, it's something that the users had prior to me joining the company, and something they were used to, so if possible I want to add this protection again.

To be protected from accidental file deletion I would recommend using built-in Volume Shadow Copy Service, pretty common practice for DR scenarios like that.

[dellock6]

If the requirement is protection against file deletion, and supposing that VM is Windows based, why not simply using Volume Shadow Copies and its "previous versions" features?
You will not be protected from VM problems since files are still inside the VM itself, but this way the load on the storage and most of all vSphere snapshots will be highly reduced. You can mix this solution with Veeam and for example take usual Veeam backups twice per day, like at 12 and 16, plus the night backup.

In this way, usual deletions can be restore to the previous seconds, major damages to servers will give you at most 4 hours RPO.

Luca.

[tsightler]

I agree with Vitaliy, using the built in Volume Shadow Copy capabilities is the far more common practice for protecting from accidental deletion and allows users to recover from their own mistakes right from their Windows desktops (or of course administrators as well).

Still, if you want to use Veeam to do this then I might suggest running two separate jobs, one that runs nightly and keeps 30 days, and another that runs every 30 minutes and keeps 20 restores points, then configure the "backup window" so that the second job only runs during the 8AM-6PM window. Both of these jobs should use the "reverse incremental" method to keep granular rollback points. The two jobs will not conflict with each other and both will be able to use CBT to only capture changes. Simple, and no powershell or offloading required.

[dsimpkins]

thanks guys - i really appreciate all the responses and ideas, this is exactly the assistance i was after.
I'll look into volume shadow copy (as per Vitaly and Luca), but also at Tom's idea of the two jobs, and Luca's idea of reducing the RPO from 8 bus. hrs to 4.

Thanks again.

Daniel