GFS Retention Policy

[aquegles]

Hello again friends.

After 20 days of setting the backup copy job, it has not presented problemas.

However, i cant see the weekly backups, and i´m wondering if i have something wrong with the configuration.

My gfs retention policy is : 30 restore points to keep, 4 weekly backups, 12 monthly backups, and 1 yearly backup.

But when i check the files of the bakcup copy job, i see only one full backup, and the rest incremental backups.

The weekly backups have a different extension, or maybe some of the incremental are the weekly backups? How can i know which files are the weekly backups?

Thanks a lot!

[veremin]

Hi Alejandro,

You will have to wait, until the latest restore point (full backup) will reach the day specified for GFS restore point creation. Kindly, check the corresponding section of our User Guide; should clarify the situation for you.

Thanks.

[aquegles]

Eremin, thanks for the response. I still don't understand how this works...

My regular backup is incremental with synthetic fulls and transformig previous fulls to rollbacks.
And my GFS retention is set to create weekly backups every sunday at 10 am.

Shouldn't already exist some weekly backups? since the backup copy job have been running for the last 20 days..

Thanks!

[veremin]

Keep in mind that GFS restore points are created from the oldest restore point in backup copy job chain (full backup, .vbk). In other words, once the specified retention is reached (in your case it's 30), the full backup will start moving forward. That's process is called transformation. When the full backup reaches the day specified for GFS restore point creation, the GFS point will be created out of it.

Thus, you should wait first - till there are 30 points in backup copy job chain, and second - till the time the full backup is moved to the corresponding day.

Thanks.

[aquegles]

Thanks again Eremin!

So, i need to wait 'till the backups reach 30 restore points.

When it gets there, it will create the weekly backups and the monthly backup? Because of my configuration..

Thanks!

[foggy]

When it gets there, you need to wait until the full backup moves onto the Sunday, then the next day the weekly backup will be created based on this Sunday's full. The same scheme applies to the monthly points - it will be created the next day the full backup reaches the day specified for the monthly backup.

[aquegles]

Ok, now i understand. Yet, i have one doubt :

This monthly backup will contain all the restore points i kept from the past month (30 in total)?

[foggy]

This monthly backup will contain a single restore point as of the day it was created.

[aquegles]

Ok, so it's not useful for my historical backups, because i need to keep 6 months of historical backups for my vm..

I will have to make backup copies and manually transfer them to tape or another storage...

[foggy]

If you need to keep 6 months of daily backups, then you have to set the corresponding regular retention setting (180 restore points) or offload them periodically.

[aquegles]

Ok Foggy, thanks a lot!!!!!!!

[tfloor]

Hi,

I'm using backup copy with GFS.
Using the following settings:

Copy every day starting at 04:00
Restore points to keep: 720
Keep the following restore points for archival purposes
- weekly 0
- monthly 0
- quarterly 4
- yearly 1

I'm now running the job since 4 july and i still see only the first full (4 july) and until today only increments.
So i expect to see a full backup somewhere in the time. But not.

Do i miss something?
Thanks

[foggy]

Tristan, you need to wait until the full backup reaches the day specified for GFS restore point creation. Quarterly full will be created right on the next day. Please see the corresponding user guide section for details.

[tfloor]

I thought i have waited long enough.

As the job begin on 4 july, then a quarter later is 5 october (it's set on First sunday of the quarter)... But there was no full... SO when do i expect anything?
ps. is it a quarter from the beginning of the job. Or a quarter from january.... ?

[foggy]

You should lower the number of regular restore points to keep. I can hardly explain this better than Vladimir did above for the weekly restore points creation:

Keep in mind that GFS restore points are created from the oldest restore point in backup copy job chain (full backup, .vbk). In other words, once the specified retention is reached (in your case it's 30), the full backup will start moving forward. That's process is called transformation. When the full backup reaches the day specified for GFS restore point creation, the GFS point will be created out of it.

Thus, you should wait first - till there are 30 points in backup copy job chain, and second - till the time the full backup is moved to the corresponding day.

Currently, the full restore point (the first and oldest restore point in the chain) is dated by July, 4th and you have something like 94 daily restore points on disk. Once you have 720 restore points, it will start moving forward each day. You need to wait until the full backup moves onto the first Sunday of the next quarter, when it is merged with the restore point created on that day. The next day the quarterly restore point will be created.

[tfloor]

Foggy,

I have indeed lower the number of regular restore points and now it is running a transforming full backup process.
I will now wait and see how the process will work the next months. But finally see a full backup creation

Thanks

[foggy]

Ok, glad we've finally sorted this. Feel free to ask if anything still not comes clear.

[TomPioreck]

I hope this isn't hijacking the thread, but I'm trying to get a firm handle on the use of disc space and how much is needed in our "live" environment. By all means, if this thread needs to be moved or modified as a new thread, I hope I haven't done anything in poor action.

Real quick background, we're using GFS for some of our servers currently, hoping to move to the system for everything. We are looking to have weekly backups for the current month, a monthly backup for each month in the last 12, and a yearly backup to be held for seven years. I'm comfortable with our retention points for the backup copy jobs and archive period setting for each period.

My issue is with the space needed. I have four weeks of weekly backup copies, 3 months of monthly backup copies, and seven days of daily backup copies. By my count, that's seven full size backup files of the server, our dailies are incrementals, not rolled up into a full each day, that's why they aren't included.

If I want to keep space for other jobs, as well as running weekly and monthly backup copy job runs, is it possible to remove completed monthly jobs from the storage area to an off-site location that's not connected, without compromising the ongoing yearly roll-up or do I only need the latest monthly job to be converted into an annual and the latest weekly job to be created for the monthly?

Thanks.

[foggy]

Tom, you do not need previous GFS restore points to create subsequent ones, so you can safely offload them to another location. Just keep in mind that corresponding records will still be displayed in Veeam B&R console until wiped out from the database according to retention.

[msc88]

Hello everybody

Im working with Veeam Backup and Replication for the first time and Im having trouble unterstanding how to configure the archive job.
We have a Backup Job which runs daily and is set to 14 Restore Points. Active full backup is set to monthly.
This seems to be fine so far.

What Im trying to do for the archive Job:

There should be 4 weekly restore points for the month previous to the current month and 1 monthly restore point for the month before this. After that there are only yearly backups.

Does this work like this:

Restore Points to keep: Does this affect all weekly/monthly etc. restore points? How can i calculate what amount is needed?

Weekly backup: 4
Monthly backup: 2
Yearly Backup: 1


Thanks in advance
msc88

[Shestakov]

Hello msc88,
As I understood your configuration reflects your needs. Just make Yearly Backup: >1 to keep more than one yearly backup.

How can i calculate what amount is needed?

Note that weekly, monthly and early backups are full backups and calculate how much capacity is going to be used.
Also, you please review the topic to have more understanding how GFS works.
Thank you.

[foggy]

Restore Points to keep: Does this affect all weekly/monthly etc. restore points?

The number of restore points to keep on disk affects only the day when the GFS restore point is created, the number of GFS restore points is controlled by the corresponding settings below.

[namiko78]

I have a backup copy job with 7 day interval, 2 restore points, and weekly GFS setup. Over the past 2 sundays, they day i expected it to generate a weekly full, it has failed to do so. No indication why. I've just changed it to start today, both before and after the current time, and changed the interval to 1 day, yet it still won't start a new GFS transformation. Am i doing something wrong?

My source is only backed up once a week, if that is a factor.

[foggy]

Steven, basically, you need to wait until the latest restore point in the chain (full backup) reaches the day specified for GFS restore point creation. Please review this thread for details.

[guitarfish]

I've read the v8 User Guide on this a few times, I just need more help.

Presently my on-site backups use a Simple Retention Policy. I run a Full backup once per month, and an Incremental daily for rest of the month. My retention policy keeps 365 restore points, so theoretically I end up with 12 Full backups per year, along with around 30 daily backups for each month.

I am now adding Cloud Connect into the picture, and it isn't feasible to store this way off-site. I need to configure the GFS Retention Policy for Backup Copy Jobs. I am focused on this screen:


What I would like to have at the end of a year:
30 days of daily restore points, one for each of the previous 30 days.
Weekly restore points for the previous 12 weeks
Monthly restore points for the previous 12 months
1 Yearly

After running this way for a full calendar year, this is what I'd want to see on Jan 1:
December – a restore point for every day, 30 total
Oct, Nov – weekly restore points, 8 total
Jan – Sept – a monthly restore point, 9 total

How do I configure the screen above to achieve this?
And while at it, Does this make sense?

[foggy]

Makes perfect sense, however "what you would like to have at the end of a year" does not 100% correspond to "what you'd want to see on Jan 1" (different number of weekly and monthly backups). Basically, you need to set retention of 30 restore points and the desired number of weeklies, monthlies and yearly backups in the corresponding fields.

You can also find more help in the thread above. Thanks.

[Shestakov]

Hello,

For the case you`ve described, the numbers should be:
Restore points to keep = 30
Weekly backup = 12
Monthly backup = 12
Quarterly backup = 0
Yearly backup = 1

Same full backup can be marked as weekly, monthly, quarterly and/or yearly, so you will not have duplications.
Also, please read the topic where your post has been merged and ask additional questions if you have any. Thanks!

[guitarfish]

OK, so if I follow, "restore points to keep" is dailies. 30 restore points gives me a month of dailies, then on top of that I have the weekly, monthly and yearly. The confusion for me is it seems like I'd have more 30 dailes, plus the additional weekly and monthly and yearly, which might come out to something like a total of 48 restore points.

Thanks much for the feedback folks. And I did read this whole thread before I started mine, FWIW.

[Shestakov]

I see where your confusing is coming from.
But retention points after the checkbox are related to GFS retention policy and not count to the 30 restore points to keep.
Thanks!

[guitarfish]

I will be bringing a Cloud Connect repository online in the next day or so, and I need to setup the Backup Copy job in v8. I have two questions:

Virtual Machines selection - At the Virtual Machines screen, I’m not clear on the differences between From Infrastructure, From Backup, and From Jobs. Are these just different methods for getting to the VM? If you select the same VM regardless of method, is it the same outcome/result?

Existing Restore Points - When the Backup Copy job runs for the first time, will it sync all the available local restore points to the cloud, while observing the defined GFS parameters? I have a Full backup and a couple weeks of daily Incrementals, all of which I’d like to get uploaded to the cloud repository. It is about 60GB, and I expect it will take about three days to upload.