Son-Father-Grandfather rotation suggestion

[mgambetti]

hi,
i just was curious about what should be the best scheduling settings to allow me to have the last 6 days/week,4 weekend/month, 12 versions/year with veeam backup.
maybe create an incremental every day of the week with retention 6 days, another backup schedule running every weekend for the last 4 weeks, and another schedule running only the last day of every month for 12 months?
tnx in advice.
marcello

[UPDATE] GFS retention has been added to version 7.0 as a part of Backup Copy jobs.

[Gostev]

Hi Marcello, that would work.

[Bunce]

Restore point option suggestion

[mckenzieaj]

Hello,

Add on question.. you can specify the "last day of the month" in a backup job? I thought the closest you can come is the last (eg) Sunday of the month!

Cheers
Andrew

[axelklos]

Hi,

I will do following:

- keep daily backups for 2 weeks
- keep latest weekly backup of month for 12 months
- keep latest monthly backup for one year


would be this possible with a workaround?

regards

axel

[Vitaliy S.]

Andrew, there is no option like that, but you can use Windows Task Scheduler to start your job via PowerShell script or simple batch file on the last day of the month.

[Vitaliy S.]

Axel, yes that would be possible, but you would need to configure separate jobs for each retention policy configuration.

[mckenzieaj]

Thanks again... and again I learn something

(mainly to read the Summary screen on the backup job config a bit better!, eg

Command Line: "C:\Program Files\Veeam\Backup and Replication\Veeam.Backup.Manager.exe" backup f4f851b4-b135-412e-9d1e-85c09ce9075a")

AM

[bbricker]

I am a new Veeam user (just bought 8 sockets of Veeam Standard) so please forgive me if this has been answered before. I'm trying to setup a schedule similar to how I've done it in other backup products. I'm not seeing how to do it without setting up multiple jobs and selecting different types of schedules (one to set weekly days, one set monthly days, one setup with explicit days, etc). But my understanding from talking to tech support is that creating multiple jobs for the same VM's will create multiple backup folder destinations, which will create multiple VBK/VRB sets, which would be a waste of space and lessen my deduplication. Is that correct?

Basically I want to be able to backup certain VM's or groups of VM's like this (for example):

Weekly: keep last 5 days of restore points
Monthly: keep only the last 4 Sundays of restore points
Yearly: keep only 1 restore point per month, for only the last 6 months
Forever: keep only 1 restore point per December (end of year copy)

What is the best way to accomplish this? The only thing I could think of was to have a script that goes into the backup folder and deletes certain VRB's from different dates. That wouldn't seem very clean, but my understanding is that I can delete any VRB and as long as I have the VBK. I would still be able to go back to a restore point of any other VRB. How are other people doing this? With our current old school backup software that is based on virtual tape disks, you just setup multiple schedules to accomplish this.

We just simply have no need to go back to infinite days over X period to do restores when 99.9% of the time we would only want to restore from the most recent backup (for example, DB servers with tons of new data added per hour). But for compliance I need to at least have SOME restore points over that X period of days (or 6 months, a year, etc., you get my point)

thanks!
Ben

[bbricker]

(sorry, I didn't search well apparently)

I see 2 different answers from previous posts on this topic that say to setup different jobs with different schedules. So to answer my question, will doing that create duplicate data in different backup folders with extra VBK's (if fulls, that's bad - a waste) and extra VRB's (not so bad I guess)?

[Vitaliy S.]

Ben, each job uses its own backup files, so if you create two jobs, you will have two chains of VBK + VRB/VIB files.

[gvinpin]

Interesting...
Last time I have asked the same question, my post was closed and I was asked to use a search.

[bbricker]

Ben, each job uses its own backup files, so if you create two jobs, you will have two chains of VBK + VRB/VIB files.

Thanks for clearing that up Vitaliy. This seems really odd to me because it doesn't look like I am the only person that wants a schedule like this. And it sounds like it wouldn't be optimal to create multiple jobs because you'd just create a bunch of duplicate VBK data.

How about my other question- could a scripted task be setup to go into the backup folder for that job and delete certain old VRB's, but leaving others to effectively create the rotation desired? I'm not even sure how I would go about doing that, and it would be quite an irritation after just spending good money on an enterprise backup product

Surely there is a simpler solution to this that people have come up with that is tried and true.

[Vitaliy S.]

Last time I have asked the same question, my post was closed and I was asked to use a search.

This time forum moderator had enough time to search it for Ben

How about my other question- could a scripted task be setup to go into the backup folder for that job and delete certain old VRB's, but leaving others to effectively create the rotation desired? I'm not even sure how I would go about doing that, and it would be quite an irritation after just spending good money on an enterprise backup product

That's not going to work as you cannot delete random VRB files from the backup chain, but thanks for the feedback.

[bbricker]

That's not going to work as you cannot delete random VRB files from the backup chain, but thanks for the feedback.

Thanks again Vitaliy, but that's what tech support recommended I could do when I called in yesterday afternoon (didn't get a case number, talked to "John"). Also- I am talking about reverse incremental, not regular incremental. Does that make any difference? I'm about to test this on some test vm's.

Again, I'm sorry to be a broken record, but I just can't imagine there is not an efficient way of doing this. With the many thousands of customers of VB&R, I would think enough people would have asked for this that there would be a smooth solution. Maybe I am just looking at this all wrong? I would love to hear feedback from others who have various retention period requirements in their environment.

[dellock6]

Incremental + syntethics + transforms do not work really well if you do not keep a full chain of backups, I think having online repositories and no offloading to tape, we all need to think in a different way when using Veeam or other software that do backup to disk.

For us, a viable workaround is to script the copy of a full vbk somewhere else with a GFS-like schedule, and leave all the veeam files in the backup repository.

[Threonine]

bbricker, I would have to agree with others here that creating multiple Veeam jobs is the best solution to add backup depth. You have an excellent point -- that backup depth is a critical consideration when it comes to a backup strategy and one Veeam probably doesn't address all that well. However, I've never seen a backup application that makes GFS rotations simple either. Those that try to make it simple are probably over-simplifying it to the point where it's not clear what it happening or flexible enough to manage. I think you'll find setting up several jobs will easily give you the flexibility and backup depth you need. Yes, of course, there is overhead in terms of disk space requirements, but you're going to get efficient restores from each job and that's far more important than deduplicating data across all your backups. It's also going to take a little more work to manage your jobs, but in the scheme of things, Veeam makes it very easy to update your job settings and understand what's going on with your backups.

[dellock6]

I think a good GFS backup schema cannot avoid using full backups for every step of the rotation.
Keeping for example forever incremental inside this kind of jobs is pretty useless to me, and the more we extend the rotation period, the more the full backups are a good way.
Think about a monthly backup, if the VM is changing fair amounts of blocks, we will probably end up having most of them changed, so it's faster to copy the full VM rather than spending time to check CBT. Also, full backups are not chained to other incrementals or reverse incrementals, and we can move the full backup anywhere without breaking the backup chain.

[frank67de]

Hello

I've read a lot of articles and I'm so confused...may be anyone can help me.

We have 15 VM machines in 5 Veeam backup jobs. We'd like to backup every day a reverse incremental and every month one full backup. Monthly backups should never deleted from the backup job !
How can I do this in one job or have I configure 2 jobs (one daily with 31 restore points and one monthly with 12 restore points)?

Thanks and best regards
Frank

Forgive me if there another topic with this...

[foggy]

Frank, you could also run one reversed incremental job and use some scheduled script to monthly offload the latest backup (which is, obviously, full with reversed incremental mode) to store it in some other location with required retention.

[Vitaliy S.]

Thanks again Vitaliy, but that's what tech support recommended I could do when I called in yesterday afternoon (didn't get a case number, talked to "John"). Also- I am talking about reverse incremental, not regular incremental. Does that make any difference? I'm about to test this on some test vm's.

I haven't tried that, but even if this scripted task deletes actual files you will still have corresponding restore point records in SQL database.

So when you attempt to recover a VM, you will have a bunch of restore points to choose from and not all of them will give you a possibility to restore the VM since there will be a "clean up" script running which will be deleting VRB files but not the SQL database entries.

[gvinpin]

If you are running Deduplication appliance, it is even easier.
I have 3 month worth of data, and then a monthly and yearly snapshots on DD appliance.

[bbricker]

Thanks everyone for the comments. I have setup multiple jobs and I think it will work out, but I'm going to have to add storage. It really does suck that I'm losing all the deduplication from one job to another. One of my groups of servers (just 3 vm's) that needs to have 3 jobs to get my restore points how I want them is using about 1TB of data for the initial full vbk write. After that, the incrementals are just 50-100GB a day. So right off the line I waste an additional 2TB by having to create the 2 extra jobs to have the flexibility of my backup scheme. Please Veeam, add some more flexible restore point scheduling! Okay, done with that rant, you guys are probably sick of hearing from me on this

Now for an odd question as a result of setting up these jobs. Do the Veeam developers just work so hard that they consider Saturday to be a "week day"?? I setup a couple of my jobs to run just on week days at 10pm, and noticed that they ran on Saturday as well. I guess I'm going to have to change it to specific days and manually pick M-F only. Possible bug?

[foggy]

Do you possibly have synthetic full scheduled on Saturday?

[bbricker]

Do you possibly have synthetic full scheduled on Saturday?

Heh, I just saw that in the patch thread and am getting the latest patch now. I have never turned on synthetics on these jobs, but I do see that saturday is the default option (even though it is grayed out). Hopefully the patch does the trick

[Gostev]

Do the Veeam developers just work so hard that they consider Saturday to be a "week day"??

And that too

[upietz]

Hello,

My question concerning the retention policy of active full backups: where is it defined?
We're setting up B&R in our VMware environment and chose to use the reverse incremental method. We are required to additionally perform monthly full backups which have to be stored for 6 months. When I'm setting up jobs, I can activate and schedule full active backups (say on first sunday each month), but I don't seem to find an option as for how long B&R keeps these backups... what am I missing?

Any help appreciated!

stefan...

[dellock6]

If you are going with the reverse incremental, is a waste to run active full again and consume other I/O on the storages. Since the most recent backup of a reverse incremental is always a full VBK file, your faster way is to pick that file after Veeam has finished creating it, and copy it in another place for your long retention needs.

Luca.

[steait]

hi guys,

trying to figure out the most optimal way to backup our vm environment.

got about 360gigs compressed data (full) and 30gigs (incremental) on-site and a 80mbit connection to a dr-site

need to keep a week worth of backup locally for fast restore - need to keep 3yrs on the dr-site

requirements:

local restore ability one week back - daily increments (weekdays)
remote restore ability 3yrs back - daily increments for a month, weekly increments for the next two months - monthly for the next 9 months - then 4 more with 6 months between them. all together 3yrs.

I cannot figure out how to achieve this best with veeam in a smart way - any suggestions on how to do this in a simple way?

[bbricker]

Steait- welcome to the club. Many of us are hoping Veeam will implement more granular retention on a single job where you can select exactly how you describe. In the meantime, read all the above threads that your post has been merged into, and it will give you some ideas. I am in the process of going away from having multiple jobs with different schedules and retention periods (which wastes time and storage from a loss of having that consolidated deduplication), and I am moving to where we have less jobs and will just manually (or script more likely) move VBK fulls to another folder for the retention that I want. Good luck and hopefully Veeam is listening to all these "feature requests"