Veeam 6.5 and backup jobs logic

[isolo]

Is there any way to create separate VBK for each VM without creating 80 separate jobs manually?
here is what I am up against:
I have 80 VMs for a total of 1.5TB.
If I need to restore a single VM from tape, I have to restore 1.5TB VBK and only then restore my VM.

[Vitaliy S.]

Hello Igor,

No, that's not possible. I'm afraid you need to create separate backup jobs to achieve this.

Thank you!

[isolo]

In that case is there a way to automate job creation? Instead of me seating and creating 80 backup jobs manually?

[tsightler]

You can use Powershell to create the jobs. There's some examples in the Powershell forum for creating a job and setting job options and it'd would be pretty easy to grab a list of all VMs and loop through to make a job for each one.

[rimvydukas]

Hi,

I'm new to Veeam so I have a question:) I created new backup job. This job archives whole esx server's VMs. Great. I archived all VMs which were on this server. I looked at the backup repository and saw that the new folder was created for this job. Initially I thought that I'll see several vbk files, each for particular VM. But as I saw - all VMs were placed in one vbk file. So I'm interested - why is that? What is the logic behind such behavior? I plan to move my archives to tape and I can imagine how long will it take to restore old full backup with a lot of VMs in it

[veremin]

In fact, it’s totally expected. However, it won’t affect in any way your ability to restore individual VM from this .vbk file - when the time comes, all you need to is to right-click required backup file and select those VMs you’re willing to restore.

If you’re eager to get separate backup file (one file per VM) then, you need to create backup jobs accordingly (one job per VM).

Anyway, I strongly recommend you reviewing corresponding section of User Guide (p.211). Thanks.

[rimvydukas]

But I want to know why did you decide to create one big full backup file which holds all VMs full backups and not separate vbk files for each machine (in job folder). I'm asking because of this - Imagine the following situation. I have one ESX server which holds 15 VMs. Each of them consumes 100GB of used space. Can you imagine total amount of vbk file if I choose to archive this server's vms? And because of that the file is so huge it will take a lot of time to transfer this file to tape and especially restore it from tape if I'll need older full backup to restore. I plan to keep only few newest full backups on disks and all old backups will be transferred to Tivoli. So I wanted to know was there any particular reason to put all VMs backups to one large file instead separating them amongst smaller files - one vbk full backup file for each VM. I do not want to create backup jobs for each VM as I think I'll get complexity, loose deduplication and so on.

As for p.211 - it contains info about Sure Backup session reports, how does this relate to my question?

[veremin]

Dedup consideration apart, but when you’re backing up folder using agent-based solution, do you expect to get separate backup file for each sub-file or just one, instead? Or when, for instance, you’re using 7zip-utility to archive, say, 15 files, do you expect to get 15 different .zip files or just one?

The same logic is here.

As for p.211 - it contains info about Sure Backup session reports, how does this relate to my question?

Sorry for that. It appears that I’ve used obsolete version of User Guide. The section I’ve referred to is called “Performing Full VM Recovery” and might be found at the page 220 in the new edition of User Guide; it might be useful if you’re eager to find information about how VM restore can be performed.

Thanks.

[foggy]

So I wanted to know was there any particular reason to put all VMs backups to one large file instead separating them amongst smaller files - one vbk full backup file for each VM.

Deduplication is that particular reason. Having separate backup file for each VM means that you lose deduplication within the job resulting in significantly more space required to store your backups. Common practice is to group similar VMs together in one job to get better dedupe among them.

[rimvydukas]

foggy,

I suspected that this is main reason that stands behind such behavior, thanks for confirming this. One question regarding deduplication - lets imagine that I have three W2k8R2/SQL 2012 servers which were prepared from one image. What deduplication ratio can I expect if I'll archive these three machines with one job. I've got only 1.2x ratio and don't know is this is expected or not?

And another thing, maybe feature request. Is it possible to introduce a feature that if I disable deduplication - program does full VMs backups not to one large file but to smaller ones (1 VM = 1 vbk file). Because sometimes the time which I need to restore old backup from tape is more important than consumed space on tapes.

[foggy]

I suspected that this is main reason that stands behind such behavior, thanks for confirming this. One question regarding deduplication - lets imagine that I have three W2k8R2/SQL 2012 servers which were prepared from one image. What deduplication ratio can I expect if I'll archive these three machines with one job. I've got only 1.2x ratio and don't know is this is expected or not?

Even though being created from a single template, SQL servers generally use small blocks and perform very small unique changes across the whole VM disk, so only blocks belonging to the OS (which are the same for these VMs) are to be deduped. I would say that 20% dedupe you have for this job is expected.

And another thing, maybe feature request. Is it possible to introduce a feature that if I disable deduplication - program does full VMs backups not to one large file but to smaller ones (1 VM = 1 vbk file). Because sometimes the time which I need to restore old backup from tape is more important than consumed space on tapes.

You could backup those VMs in a separate jobs, I suppose.

[rimvydukas]

Yes, I could create a backup job for one/two/three VMs, but if I have 50 VMs and want to have smaller full backup files while sacrificing deduplication - such option would be more than welcomed. I do not want to have additional 50 jobs at all:( It would be so time consuming to manage all these backup jobs.

[foggy]

Agree, this is not the case for the large number of VMs.

[rimvydukas]

So, I think, I'll leave this as feature request for your consideration:)

[foggy]

Sure, thanks for the feedback.

[flierville]

Hello everyone,

Is there an option in the JOB that allows to have one vbk VM file?

I have about 80 VM to save, so I want to avoid making a JOB VM.

I am well aware that this action would consume disk space but personally I find that making a single large file that contains all the VM is dangerous (if disk crash) and difficult to handle (copy on Nas etc.).

Thank you in advance for your return.

[foggy]

There is no such an option, please see considerations above. Thanks.

[flierville]

I think this feature is set for an upcoming product improvement.

[Vitaliy S.]

I am well aware that this action would consume disk space but personally I find that making a single large file that contains all the VM is dangerous (if disk crash) and difficult to handle (copy on Nas etc.).

But if disk crashes, then it doesn't matter whether you have 1 file or multiple ones, most likely all of them might get corrupted. Can you please tell me if you have considered using backup copy jobs to transfer your backup files to another location? How many backup copies do you currently maintain?

[yizhar]

Hi.

I have suggested a different approach some time ago, and named it "job group" feature request.
The idea - setup several jobs in Veeam, but with common shared schedule and reports.

Veeam feature request - job group view topic
http://forums.veeam.com/viewtopic.php?f ... 93&start=0

Yizhar

[flierville]

But if disk crashes, then it doesn't matter whether you have 1 file or multiple ones, most likely all of them might get corrupted. Can you please tell me if you have considered using backup copy jobs to transfer your backup files to another location? How many backup copies do you currently maintain?

Precisely to prevent a crash, we copy our backups we outsource.

Eventually, the idea is to manipulate the VM backups (copy to tape, disk copy, copy of a VM on a client pc to launch VMware Player etc.).
In this context, it is easier to manipulate a file rather than a very large VM files to multiple VMs.

We have about 100 Vm to save.

[ceez]

Hello everyone,

Using BR7.0 to backup our vm guests. If I create a job and select our servers it creates one large backup vbk file containing all the servers under the assigned backup repository. I would like each server to have its own folder within that repository.

So right now it looks like this:

E:\backups\company backup\<VBK FILE CONTAINING ALL SERVERS.

I would like for it to look like this:

E:\backups\company backup\DC\<the DC vbk file>
E:\backups\company backup\EXCH\<the EXCH vbk file>
E:\backups\company backup\PRODUCTION\<the PRODUCTION vbk file>
etc...

The only way I can accomplish this so far is to create backup jobs for each server which can become very crowded.

Thanks,

Ceez

[foggy]

Cesar, this is totally expected, please review the thread above for details. The only way to create several VBK files is to create several backup jobs.

[ceez]

Cesar, this is totally expected, please review the thread above for details. The only way to create several VBK files is to create several backup jobs.

thanks foggy.

[gmcoburn]

Hello,

Hey, two posts in first registered day. Long time user, first time poster...

I currently back up a large number vm's within our vSphere infrastructure across multiple sites using Veeam B&R 7.0.0.871.

To ensure that new VMs are backed up, we ensure that when any new VMs are deployed they are deployed into VM folders, and we have targeted our backup jobs to the VM Folders.

Whilst the backups are progressing well, the resultant backup files are very large, and are presenting us with issues related to large backup files, their transportability, offsite storage and retention of particular individual VMs, and potential backup data loss in the case of corruption of a single large backup file that contains many VMs.

As a feature request, what I would like to propose is that we are provided an option within a backup job that is targeted to a VM Folder for it to create either individual backup jobs, or individual backup data files per VM with a single Backup Job.

I am aware that we would lose Veeam Job inline dedupe, but with the proliferation of dedupe external to Veeam, Windows 2012 Data Deduplication for example , I would be happy to accept that trade off.

If this is possible now and I've missed an option somewhere, I apologise in advance, and can you direct me to the required job setup option.

If not, then feedback is appreciated.


Regards

George

[foggy]

Thanks, George. Your request was merged into existing discussion along with similar requests, so you could review some considerations lying under the current implementation.

[gmcoburn]

Thanks foggy, I had seen these posts, and its is good to know that our needs are not unique and that Veeam is aware of the desire in the Veeam community for the functionality as stated. Here's hoping it grabs some traction.

George

[veremin]

Yes, these forums are thoroughly monitored for potential features/improvements. So, you've chosen absolutely right place to share your feedback. Thank you!

[Wocka]

Hello,

I'm trialling Veeam to see if it suits our comany's needs. So far it's looking good.

I have a question that I can't seem to work out. When I create a Backup Job that has multiple virtual servers included, my backup file name is that of my job name (eg Test Environment2014-10-01-xxxxxxx.vbk) and not that of my servers (eg server12014-10-01-xxxxx.vbk & server22014-10-01-xxxxx.vbk). Why does it combine the 2 servers into 1 backup file?

Should I create jobs for groups of servers or should I create a Backup Job for each server? What is the benefit of grouping servers together, especially if I would like to replicate across a WAN or create File Copy jobs.

[foggy]

Warwick, what you're observing is totally expected and was designed to be so, due to considerations described above. You can also search for other existing topics discussing recommendations on grouping VMs into jobs. Thanks.