[ARCHIVED] Veeam Backup & Removable Media (rotated drives)

[foggy]

Andrew, I've merged you post into this existing discussion as it seems you are looking for the implementation of the similar scenario. Please review this thread and feel free to ask any additional questions, should they arise. Thanks.

[jroback]

We've had good luck using the high-rely 2 & 4 bay netswap plus for this (http://www.high-rely.com/hr_66/products ... swap-plus/) . These are iscsi targets that have an integrated mirroring raid controller, so VEEAM sees etiher one or two 4 TB Volumes, and the device automatically raids each 4 TB drive to another 4TB drive in realtime. We configure Veeam to backup to the 4 TB volumes as usual, with as much retention as space allows, and then we have the client eject the mirrored 4 TB drive as often as they like to send data offsite. This is a nice easy to maintain solution because the "Main" backup is alway onsite, and the offsite backup is a full-self-contained backup as well.

[Damian Netspeed]

I want to rotate 2 external hard drives for offsite backups.
What is the best way to setup a veeam backup job for this ?
Can I do it with a single job or do I need to create 2 jobs ?

[foggy]

Damian, single job is enough. Please see recommendations given in this thread.

[Michalis]

#case number is 00517297#

Hello,
we've purchased veeam backup & replication, version 7.0.0.764. I'am pretty new in that, so this may sound trivial, although i didn't find any relative posts in the forum. The problem, as the titles implies, is that veaam coppies the new .vbk files into the day's folder instead of replacing the old one which results in mail warnings of getting low in disk space. We are using 2 rotated external drives and there's no incremental backup policy applied. The files are been saved under the name "Friday2014-01-17T230037.vbk" for example nad there's also the relative .vbm there. Could would ask for some help with that? p.s. : Wouldn't be more appropriate to apply incremental backup?

Thank you in advance, Michalis

[foggy]

Michalis, please review the thread above for instructions on how to implement a rotated drives scenario. Basically, you need to specify a couple of registry values. Be sure to update to the latest Patch 3, as it adds full support for this scenario (registry values can be found in the patch release notes).

[Morgenstern]

Hey guys,

I was hoping if you could advise me on the best way to carry out daily backups on the 1TB RDX cartridges.

I have five of them, one for each day. The total size of a full backup is roughly 600GB and my idea is to have that incrementaly updated once a week. For example the Monday cartridge gets an initial full, then is taken off site and brought back following Monday to be updated, then taken away again.

I need the size of this specific backup remain the same, i.e. not to grow every time a new backup cycle is run. I have set it to the Incremental with synthetic fulls (forever incremental + Transforming previous full backup chains into rollbacks). Restore points to be kept on disk are set to one.

I imagine it as a comressed copy of all vm's disks that is incrementally updated/overwritten with changes at each new cycle without any increase in size. Does it work like that?

I have other jobs designated to do proper weekly backups that aren't limited by the storage space. These RDX cartridges are only meant for keeping track of the most recent changes which are then backed up elsewhere anyway.

Many thanks,

Morgenstern

[foggy]

Martin, please review the thread above for instructions on implementing the rotated drives scenario. With a single job you will be able to have a new full on each cartridge every day. If you need incrementally updated full backup on each drive, you can create several jobs, each writing exclusively on its own drive, and schedule them to run on corresponding days.

I need the size of this specific backup remain the same, i.e. not to grow every time a new backup cycle is run.

VBK file will most likely grow, since the new data will occupy new blocks rather than replace existing ones.

I have set it to the Incremental with synthetic fulls (forever incremental + Transforming previous full backup chains into rollbacks). Restore points to be kept on disk are set to one.

No need to enable transform here as the chain will contain a single restore point only.

[Morgenstern]

If you need incrementally updated full backup on each drive, you can create several jobs, each writing exclusively on its own drive, and schedule them to run on corresponding days.

I am already doing this. I have a separate job set up for each RDX cartridge.

VBK file will most likely grow, since the new data will occupy new blocks rather than replace existing ones.

So is there no way of getting veeam to delete the older data to avoid going over the storage capacity? It seems that this way, when the disk runs out of space I will have to create a new backup job, because the old one won't work if I delete any of the old data manually.

[foggy]

With Veeam B&R v7 patch 3, you can configure (via registry) automatic cleanup of the drive before writing to it, however in this case active full backup will be performed during each job run (not incremental, as you're asking for).

However, what I'm talking about is natural growth of the backup file due to the new data generated inside VM guest OS (unless it is a pretty static VM, not producing any considerable new data).

[Morgenstern]

With Veeam B&R v7 patch 3, you can configure (via registry) automatic cleanup of the drive before writing to it, however in this case active full backup will be performed during each job run (not incremental, as you're asking for).

However, what I'm talking about is natural growth of the backup file due to the new data generated inside VM guest OS (unless it is a pretty static VM, not producing any considerable new data).

I can live with having a full backup instead of incremental one. I will end up having a backup for each day anyway. So do I just install the patch and rectreate all jobs from scratch? Or do I keep the existing ones with the forever incremental + full backup + 1 restore point scheme?

[foggy]

You can leave you current jobs, however be sure to specify the registry keys described in the patch release notes and restart the Veeam B&R console. Also, with the said registry modification you can go with just a single job, it will detect the drive change, cleanup the repository, and perform full backup.

[Morgenstern]

Thanks a lot, I'll give that a try!

[Morgenstern]

You can leave you current jobs, however be sure to specify the registry keys described in the patch release notes and restart the Veeam B&R console. Also, with the said registry modification you can go with just a single job, it will detect the drive change, cleanup the repository, and perform full backup.

Sorry to sound a bit dull, but do those registry keys get created and enabled automatically when I install the patch? Or do I need to use regedit to create them afterwards? If the latter is the case, could you please give me the full path for both?

[foggy]

You should create those keys manually under HKEY_LOCAL_MACHINE\SOFTWARE\Veeam\Veeam Backup and Replication hive.

[Morgenstern]

Thanks a lot!

[Erik]

I updated Registry keys, created a new folder on a new drive shared with the same name and drive letter and the process seems to work. I notice that when I swap drives, it creates a new VBK file (Doing reverse incremental) - which is not a surprise. However, when I go to restore a VM, it only shows me one restore point, when the other drives had 13 days of backups on it? Am I doing something wrong?

[Vitaliy S.]

Hi Erik,

You're doing it right. You don't see older restore points, because the storage containing those was removed. B&R cannot know if it is going to come back ever, so it cleans up the configuration database immediately. To restore VMs from older restore points, you need to plug the removed hard drive back in, and import the backup files to get their contents appear as available restore points.

Thanks!

[veremin]

Hi, Erik,

Do the drives in question have any distinguishing factor, like ID, internal name, etc., using which you can separate them from to the other? I'm wondering, since you can, probably, write a custom PowerShell script in order to have a required visibility across restore points.

Specified as post job activity, the script will track the job end time, as well as, the ID of currently inserted drive and, then, write corresponding information to the specified file. (Something like, "Restore point created at 7/16/2013 20:05:12 resides on Drive #12309").

Thanks.

[Morgenstern]

Hey guys,

I am wondering if you could help me out with some details on my daily backups on rotated RDX drives. After installing and setting up the latest patch, the job I have set up creates a new full every night. This job is set to only hold one restore point.I thought this would be fine for restoring stuff from any day of the past week. Unfortunately the restoring wizard only sees the most recent restore point written into the repository info and is able to restore only from the latest backup. If I insert any of the other four RDX drives I have, the restoration wizard fails with the following error:

Code: Select all

04/03/2014 10:05:56          Enabling backup operator privilege
04/03/2014 10:05:56          Starting file level restore
04/03/2014 10:05:56          Interactive mode
04/03/2014 10:05:56          File Level Restore started
04/03/2014 10:05:56          Mounting restore point
04/03/2014 10:05:57 Warning    File does not exist. File: [E:\RDX Daily\RDX Daily2014-03-03T183125.vbk].
                               --tr:Failed to open storage for read access. Storage: [E:\RDX Daily\RDX Daily2014-03-03T183125.vbk].
                               --tr:Cannot apply patches stored in folder [HostFS://E:\RDX Daily\RDX Daily2014-03-03T183125.vbk[699614a2-6f47-4e58-92e4-bb185d4cfdaa (12)]].
                               --tr:Failed to build restore point.
                               Failed to restore file from local backup. VFS link: [summary.xml]. Target file: [MemFs://Tar2Text]. CHMOD mask: [0].
                               --tr:Cannot process [tar2text] command. Li
04/03/2014 10:05:57 Error    MountRestorePoint failed, mountId '8570e6fa-5c02-4f06-a1bb-4d84c43eb5ac'
                             FlrManagementService: Failed to mount restore point (mountId:8570e6fa-5c02-4f06-a1bb-4d84c43eb5ac)
                             TextFromTar failed
                             Client error: File does not exist. File: [E:\RDX Daily\RDX Daily2014-03-03T183125.vbk].
                             Failed to restore file from local backup. VFS link: [summary.xml]. Target file: [MemFs://Tar2Text]. CHMOD mask: [0].
                             
04/03/2014 10:07:16 Error    Job has failed unexpectedly

Could this be resolved by raising the number of restore points from one to five (for the five RDX drives I have)? Or perhaps by creating a separate job & own repository for each day?

[veremin]

H, Martin,

How did you perform restoration procedure? You imported the backup file to the backup console prior to restoring, then, initiate restoration procedure, and got the corresponding error, right?

Thanks.

[Morgenstern]

How did you perform restoration procedure? You imported the backup file to the backup console prior to restoring, then, initiate restoration procedure, and got the corresponding error, right?

Hi, I did not do the impoting bit prior to restoring previously. But now that I've tried it, I am getting a "Backup already exists" message and then the same as before.

[veremin]

Hmmm, this looks strange, as the references to the previous restore points should have been cleared out from configuration database.

In this case, the best option will be to open a ticket with our support team and let them confirm whether everything is setup correctly.

Thanks.

[Morgenstern]

Okay, cheers

[Morgenstern]

Hmmm, this looks strange, as the references to the previous restore points should have been cleared out from configuration database.

Resolved! When I was trying to import the last week’s backup using a .vbm file, I was getting a “Backup already exists” error message and then obviously the “MountRestorePoint failed”. But when I imported the .vbk file instead, it all went fine and then I was able to do a restore of a week old backup from an RDX drive.

[veremin]

Most likely, copied .vbm file had references to previous restore points, thus, the import process failed. Once you imported full backup directly, everything worked as expected.

Anyway, thanks for coming back and updating the topic with the resolution; much appreciated.

[dirkdigs]

I am looking to use a set of 5 external hard drives for offsite rotation of backups.

currently i have approx 10 vmware VM servers backing up to local disk.

is it possible to setup a second job and point it to external disk?

how do i keep the external usb drives in sync. is there a specific backup retention policy i should be using.

thank you.

[Vitaliy S.]

In order to make it work with backup jobs in v6.5 you need to specify the registry key mentioned on the first pages of this thread. As to the retention policy, then you should be controlling it manually. See my previous post above for more details. Hope this helps!

[dirkdigs]

What sort of preparation is needed on the external hard drive?

Am I correct assuming that as long as when each hard drive is connected to the server that they are inheriting the same Drive letter each time that it will work??

Would you create a folder for the backups or backup to the root of the drive?

[foggy]

What sort of preparation is needed on the external hard drive?

With 6.5 you should clean up old backup files prior to running the backup job.

Am I correct assuming that as long as when each hard drive is connected to the server that they are inheriting the same Drive letter each time that it will work??

Yes.

Would you create a folder for the backups or backup to the root of the drive?

It's completely up to you, you can create a repository at any suitable path.