The need for Active Full backups

[ChrisDriver]

I have a small virtualized environment (2xESXi 5.0 servers and 5xWindows VMs) and my current Veeam backup strategy is the following:
Daily - Reverse Incremental backup, 31 restore points, backed up daily
Monthly - Reverse Incremental backup, 12 restore points, backed up once a month

I have read in numerous posts that it is recommended to periodically perform an Active Full backup. I have also seen it mentioned this is more important for Forward Incremental backups than Reverse Incremental backups. It appears the sweet spot is every 1-3 months, but surely the frequency of Active Fulls in dependent upon the number of incremental backups performed within that time frame? Please can anyone offer advice regarding the recommended frequency I should perform an Active Full Backup for my Daily and Monthly Reverse Incremental jobs.

In the future I am planning to change my backup strategy to Forward Incremental Forever so if possible could you also advise me on the recommended frequency of Active Fulls for this type of backup method.

Many thanks in advance

[Shestakov]

Hello Chris,

Now, having Surebackup for recoverability testing you need to run Active Full manually:
- If SureBackup detects unrecoverable backup, or backup file data corruption
- If you want to reduce VBK size (after rearranging jobs or deleting lots of VMs)
- After major upgrade of our software to profit from new enhancements

However, if you don`t leverage Surebackup capabilities, you can also schedule the Active Full.

It appears the sweet spot is every 1-3 months, but surely the frequency of Active Fulls in dependent upon the number of incremental backups performed within that time frame?

Yes, the frequency of Active Fulls depends on the number of runs. In your case I would do active fulls once in 3 months for the daily job and approximately once a year after the Software update and/or job configuration changes for monthly job.

In the future I am planning to change my backup strategy to Forward Incremental Forever so if possible could you also advise me on the recommended frequency of Active Fulls for this type of backup method.

I would recommend to use Surebackup job and make active fulls only in above-mentioned scenarios.
Thanks!

P.S.This link can give you additional useful information. Please review.

[ChrisDriver]

Many thanks for your reply and the link.

I should have mentioned that I have the Standard edition of Veeam which unfortunately does not include SureBackup (message to Veeam: Please include SureBackup in all editions!).

As SureBackup is not available to me, please could you revisit my final question and if possible recommend the frequency of Active Fulls for Forward Increment Forever jobs.

Thanks!

[Shestakov]

I see your point, Chris,
Note that if you enable synthetic and/or active full backups, Veeam Backup & Replication will produce a forward incremental backup chain.
If you want to use Forward Increment Forever method without Surebackup, I would still recommend you not to schedule Active Full for the monthly backup job, since there is no reason to worry about recoverability of 12points-long chain. For daily backup you can schedule active full to run once a month.
Thanks!

[foggy]

Chris, please review this thread for recommendations on full backup frequency for different backup methods: How often do I need a full backup?

[ChrisDriver]

@shestakov

Note that if you enable synthetic and/or active full backups, Veeam Backup & Replication will produce a forward incremental backup chain.

I am surprised at your statement above. If my backup job is configured with a backup mode as Reverse Incremental and I also enable the Active full backup checkbox are you saying that Veeam will produce a forward incremental backup chain? I would have expected the Reverse Incremental backup chain to continue with the only differences being a) When the Active Full backup runs according to its schedule, an additional .vbk file would be produced b) Regular backups run after the Active Full would use the Active Full's .vbk file as the 'baseline' full backup to build its reverse incremental chain from.

[Shestakov]

No, I mentioned that by definition of Forward Incremental Forever, the method does not contain Full backups(neither Active nor Synthetic) and if you schedule one, Forward Incremental Forever will become a simple Forward Incremental.
Your understanding about Active Fulls in Reverse Incremental method is correct.
Thanks.

[ChrisDriver]

Thanks for your explanation, I realise now you were referring to Forward Incremental Forever backup methods!

One final question if I may. I am planning to change my backup method from Reverse Incremental to Forward Incremental Forever. The reasoning behind this is that I am planning to use Veeam Cloud Connect and I understand that forward incremental backup strategies are more efficient than Reverse Incremental when copying backups over the internet.

The Forward Incremental Forever strategy seems ideal, but as you mentioned, I cannot schedule an Active Full backup. From this can I assume that Veeam does not consider running an Active Full Backup necessary when using Forward Incremental Forever? To me this seems risky. I have read forum posts and Veeam articles discussing edge cases on how corruption can occur within Veeam backups. It appears that without SureBackup (not available in the Veeam edition) or a semi-regular Active Full Backup, this corruption could exist unnoticed and affect an entire backup chain. If this is the case how can I gain the benefits of Forward Incremental Forever backups but ensure my backup chains are OK?

Many thanks and I really appreciate your time assisting me with this.

[Shestakov]

Correct, if you want to avoid long-chaining you can just schedule simple Forward Incremental method with monthly Active Full.
If you use Surebackup or have only short chains(like your monthly backup, you can choose Forward Incremental Forever method and do Active Fulls manually when it`s needed.

By the way, we`ve got a number of clients having 100+ long chains with no issue. However, it`s up to you to decide.
Seeing your situation, I would recommend you to choose Forward Incremental Forever method for the monthly backup job and simple Forward Incremental with monthly Active Full for daily backup job.
Thanks!

[ChrisDriver]

Thanks for your advice regarding this matter... much appreciated!

[Shestakov]

You are welcome Chris,
Your questions are totally sane, so don`t hesitate asking once you have a question. Thanks!

[redhorse]

Hello,

I need some ideas for our backup design. As I understood from many posts the recommended backup method is forever forward incremental but in my case it is necessary to keep about 60-90 restore points of our file server on disk. Isn’t it a too large backup chain? We use Veeam Enterprise so I would be able to do a SureBackup job, but how often should I do that? I keep 15 restore points from the other VMs, how often is SureBackup verification necessary? These are about 30 VMs, so what is the recommended way to configure a SureBackup Job for all these machines?

Instead of using forever incremental I can use a regular forward incremental, is it sufficient to start an active full once a month and a synthetic full once a week with transforming previous restore points to reverse incremental?
In both cases I have to move at least one full backup to tape once a week and nice to have the increments during the week, is that possible in these scenarios?

Thanks and nice weekend

[Shestakov]

in my case it is necessary to keep about 60-90 restore points of our file server on disk. Isn’t it a too large backup chain?

Even longer than 60-90 long chains will be 100% reliable if you use Surebackup, but the restore can take more time.

We use Veeam Enterprise so I would be able to do a SureBackup job, but how often should I do that? I keep 15 restore points from the other VMs, how often is SureBackup verification necessary?

Depends on your needs and capabilities. I would say, once a day is perfect, once a week is good.

These are about 30 VMs, so what is the recommended way to configure a SureBackup Job for all these machines?

I would suggest to group VMs for surebackup verification based on their SLA, application or installed OS.

Instead of using forever incremental I can use a regular forward incremental, is it sufficient to start an active full once a month and a synthetic full once a week with transforming previous restore points to reverse incremental?

It`s fully described in the thread above, please take a look.

In both cases I have to move at least one full backup to tape once a week and nice to have the increments during the week, is that possible in these scenarios?

It`s possible with both scenarios, but in my opinion, the least resource-consuming solution is to schedule a source backup job to run weekly full backups just before, backup-to-tape job run.

Thanks.

[foggy]

We use Veeam Enterprise so I would be able to do a SureBackup job, but how often should I do that? I keep 15 restore points from the other VMs, how often is SureBackup verification necessary?

You'd want to test every restore point to be sure you're able to recover from it.

[Treeeman]

Hello,

maybe a stupid question but i am a little bit confused.
Veeam Support told me one or two years ago to setup a periodically Active Full Backup in Case the Backup Chain get corrupted.

A few days ago i read some Informations the Active Full is not really needed any more because the Backup Health and Integrity Checks will determine if there is a Problem with the Backup Chain.

Please clarify this issue for me.

Many Thanks!
Kind Regards
Marco

[foggy]

Marco, indeed, our recommendations regarding this have slightly changed. I'm merging your post into the thread that should answer your questions. Feel free to ask for additional clarification though, if required.

[Treeeman]

Thank you, Alexander.
I think i got it. It depends...

What i understand:
If i am not willing or able to do a surebackup it is still reccomended to plan a periodicall Active Full. Expecially if there is a long Backup Chain.

My Setup:
In my case i have about 25 VMs. I added Jobs for each Type of Windows-VM (Active Directory DCs, Exchange, File, SQL). All other Windows-Server are in one Backup Job for best dedup experience. I do a Forward Incremental Forever (without a planned Active Full) and additionally a Backup Copy Job to another fire protection area. Additionally i created a SureBackup Job for each VM once a week. So if there will be detected an error i would do a manual active full.
This Setup fine? If i read the previous post, i think so.

[foggy]

Yes, your understanding is correct.

[Shestakov]

And you setup looks fine.
More best practices about data protection, chaining etc. can be found here.
Thanks.

[Lagcat]

Hi All

we have always done reverse incremental but we are looking to change to forward speed things up for our backup window which always seems to get smaller and smaller

my question is:

if i run a synthetic every Friday - should i still schedule an active full weekly/monthly or does a synthetic do the job?

i don't want to only use synthetics if a few weeks down the line i hit problems and asked to do a active full

any advice on this - i could read through best practice but i would rather some use info back

Cheers

[Shestakov]

Hi Jake,
As described above in the thread, the best practice is to use Forever Forward incremental method with Surebackup. In this case you need to run active fulls manually in certain situations such as:
- If SureBackup detects unrecoverable backup, or backup file data corruption
- If you want to reduce VBK size (after rearranging jobs or deleting lots of VMs)
- After major upgrade of our software to profit from new enhancements
Please review the topic for more info.
Thanks!

[ChrisDriver]

As described above in the thread, the best practice is to use Forever Forward incremental method with Surebackup. In this case you need to run active fulls manually in certain situations such as:
- If SureBackup detects unrecoverable backup, or backup file data corruption
- If you want to reduce VBK size (after rearranging jobs or deleting lots of VMs)
- After major upgrade of our software to profit from new enhancements

- Or you have the standard edition of Veeam B&R which does not include Surebackup!

[veremin]

Not that good, but with Standard Edition you can still emulate SureBackup manually.

[cffit]

I was doing forever incremental backups with weekend synthetic full for some time. Then our weekend synthetic full began to take a long time to build. I was told by support a while back that we should be doing periodic normal full backups or else this can happen. Do others out there do forever synthetic fulls only?

1. Is this a true statement?
2. If so, does that still hold true with v8 today?

I am starting to do backup copies offsite and it seems like they are forever synthetic fulls, so it got me to wondering.

[Gostev]

1. Yes, due to fragmentation.
2. Yes.

[Shestakov]

1.Yes
2.Yes (don`t forget about Surebackup)
Please review the topic for additional information.

[MP230]

Hi,

while its on the table i wanna throw in another question about the active full frequency. What i am missing (if it´s there, pls tell me) is that i can choose to have an (scheduled) active full every x restore points. Since the job is counting in Restore points and not in days, while the scheduling for active full does. Would make things easier in certain cases. Is this planned for future releases?

Regards

MP

[Shestakov]

Hi,
There is not a lot of demand to schedule Full backup as each Xth restore point and it looks not difficult to calculate the number of restore points between Fulls.
Could you specify your use case, why do you need the option?
Thanks!

[MP230]

It´s not about difficulty to calculate, it´s about having to take care about it. If you for example agree to have a max. chain length of 15 (as a standard, when not utilizing surebackup) allthough you wanna keep a lot more restore points (50+), you will need to think about a lot more parameters than just one setting for having active fulls every x restore points like, actively scheduling for weekly, monthly - exclude include months, when does the job start according to included or excluded months, job runs aside from normal schedule, etc...

Regards

MP

[Shestakov]

Thanks for the explanation.
We recommend as best practice to keep long chains offsite(at least another repository) using backup copy jobs.

The backup copy job has GFS retention policy which helps to organize weekly/monthly/etc. restore points.
In this way you don`t need to have a struggle with complicated backup job scheduling.
Thanks!