v6 - How to have a local and off-site backup copy?

[Venzent]

[merged]

Hello there,
We have Veeam B&R 6 Enterprise in our datacenter, running on VmWare 5. Between the datacenter and our office is a 100 mb line.
In our office is a iSCSI storage, without any active components, so no servers. I can access the iSCSI storage from our datacenter by iSCSI.
I want a DR backup of the Veeam backup-files off-site to the office and thinking of using a robo-copy sync script for it. It's about 18GB/day so that should be possible.
I want to use WAN optimisation for the storage and best compression to keep the files as small as possible.
Is this the best way for doing this, or are there some options to let Veeam B&R write to 2 locations in 1 job, or any other option?

[foggy]

Ralph, there are plenty of other existing topics on this forum about offsite replication of backups using third-party software, please use search to find them: http://forums.veeam.com/search.php?st=0 ... te+example.

[kewnev]

For an off-site backup copy, I wonder if anyone here has looked at the open source product "Areca Backup"? I use it for backing up some of my data to an FTP site, and it works well. The advantages are:
- Supports deltas, so only changed data is copied over
- Encrypts data before sending it over - so no need to worry about having any sort of encryption technology at the destination end
- No daemon is required at the destination end
- Can store to a local drive, network drive, FTP, FTPs (with implicit and explicit SSL / TLS) or SFTP server
- Can retrieve data at any point in time
- You can easily see how much data each transfer uses
- Very easy to use GUI, but also has command line support for automation & scheduling

http://www.areca-backup.org/

When I get time I'll do some tests with Veeam backups.

[bhwong]

The basic question I will ask about any backup is,

1. Will areca-backup be able to backup locked file? If not, how does it overcome this?
2. And for backup such as database, will it be able to trigger a quiesced to ensure there won't be any partially written file that lead to corrupted file when restored?
3. Any consistency check or verification that the backup is restore-able?
4. VSS support on Windows?

[kewnev]

What I actually meant was to make an off-site backup copy of the Veeam files. Not a replacement for Veeam.

Veeam does all the backup, and then use Areca to replicate Veeam changes to an off-site location.

The basic question I will ask about any backup is,

1. Will areca-backup be able to backup locked file? If not, how does it overcome this?
2. And for backup such as database, will it be able to trigger a quiesced to ensure there won't be any partially written file that lead to corrupted file when restored?
3. Any consistency check or verification that the backup is restore-able?
4. VSS support on Windows?

[Venzent]

@kewnev
Altough it sounds good, I'm very carefull in letting some unkown program to take responsibilty in my DR-data....In case of reversed incrementals, there's only 1 full backup and how sure are you that the file is consistent?
But. I will deffinitly take a look at it.

[kewnev]

That's a fair point.

The features say "Transaction mechanism : All critical processes (such as backups or merges) are transactional. This guarantees your backups' integrity."

Whether that ensures consistency, not really sure.. all I can say is that for my (non-Veeam) uses, I have had 100% restore success.

@kewnev
Altough it sounds good, I'm very carefull in letting some unkown program to take responsibilty in my DR-data....In case of reversed incrementals, there's only 1 full backup and how sure are you that the file is consistent?
But. I will deffinitly take a look at it.

[boromic]

I have been running two jobs to backup locally and remotely. It worked pretty well for a while, but the problem is that it just takes 1 day of downtime to increase change rate and then the jobs will be out of sync and I can never make my backup windows. This is because the order Veeam processes VMs is completely random and if the Veeam VM is processing a VM in another job, the other job will be held. When the Veeam VM is the last one to be processed in the job (because it is skipped), and my other job is processing a 1 TB Linux VM the job can wait 4-9 hours before it can finish. I have about 14 VMs and the majority of them are 2008 R2 and I just process my whole vSphere container in both jobs for simplicity and to get the best de-duplication. I have a 25 Mbit WAN connection that is throttled to 15 Mbit for Veeam during normal business hours and off hours it gets the full 25 Mbit.

[rbrambley]

This is because the order Veeam processes VMs is completely random

as of v6 you can set the order that VMs are processed in the Ad VM window of a job. Of course, you have to pick individual VMs (and not a container) to take advantage of this. I guess you could create folders of VMs in vSphere as your containers and then specify the order of the folders.

if the Veeam VM is processing a VM in another job, the other job will be held. When the Veeam VM is the last one to be processed in the job (because it is skipped), and my other job is processing a 1 TB Linux VM the job can wait 4-9 hours before it can finish. I have about 14 VMs and the majority of them are 2008 R2 and I just process my whole vSphere container in both jobs for simplicity and to get the best de-duplication.

Sounds like you should consider separating the 1 TB Linux VM to it's own jobs (both for local and offsite). That way you won't hold up all of the other VMs because they will be in the current jobs (local and offsite) and "fly around" the slower job on the Veeam Proxy(s). Make sure your Proxy(s) can handle more than one concurrent task by giving them the most CPUs you can afford and adjusting their Properties > concurrent task setting.

You'll probably also benefit from creating a new Proxy dedicated for the offsite jobs. A separate proxy for local ad offsite jobs gives you more dedicated resources for each instead of overloading one for both. Use one of your existing 14 VMs or build a new one for this.

Consider the post job command option of your jobs to link your Veeam jobs too. For example, link the offsite job to start as soon as the local job ends. Then you only have to schedule the local jobs.

Just a few ideas how to meet your window because you can leverage the Veeam architecture and job design to get more done during that time.

True, you will have to compromise some dedupe to meet your window. Consider a dedupe storage device as a Veeam Repository to get global dedpe across all the different Veeam jobs. Could be as simple as Server 2012 in the near future or a device like Exagrid. Exagrid, and most other storage dedupe devices, replicate the files for you so you don't have to do it with Veeam jobs.

It's a trade off between space consumed and meeting your current window.

[lanzia]

[merged]

Hi all,

For the first time I'm in the process of confuguring Veeam Backup & Replication 6.1 and I can use a bit of help.

the Scenario:

Site 1:
All VMWare hosts and servers are located at this site,
also there is a QNAP NAS located at this site for Veeam to use as backup storage.
This site also contains a physical server wich has Veeam backup & replication 6.1 installed.
A backup repository is created to make use of a CIFS share located on the NAS in site 1

Site 2:
Site 2 is a branch office, witch no serves.
A second QNAP NAS is present at this site.

I would like to replicate the Veeam back-up data located on the NAS at site 1 to the NAS at site 2.

Is there any way this can be done by Veeam backup & replication 6.1 ?

Kind regards,

Alex

[jumbo22]

[merged]

Hello,

I wonder if anyone can confirm my thoughts on setting up local and remote site backup

The source site already has B&R running on a VM backing up to local storage. There are three hosts running ESXi5 and 20 or so VM's. Data size is

approx 2.5Tb and daily changes amount to roughly 70Gb. The resulting backup size on storage is 1Tb

I need to be able to backup the VM's to a DR site and for certain VM's retain 5 days incrementals and 4 weekly backups

My understanding is this... I need to install Veeam B&R at the D/R site as well (in the event of a catastrophe at production site) so I can

restore the backups.

I will:

Install ESXi node at D/R site for recovery of backed up VM's
Install Veeam B&R at D/R site (no extra license needed as still only backing up same number of VM's) on Windows 2008 R2 VM
Create backup repository at D/R site as well as proxy (used as destination proxy) on the Veeam server at D/R
Create a new backup proxy at the source site (used for source proxy)
Take a copy of a full backup from the source site and use to seed new backup job on Veeam at D/R site
Setup the backup schedule at D/R site not to coincide with that at the source
Use forward incrementals (less load than using reverse incrementals) and synthetic full backups (so as not to interrogate the VM's at source)
Use Instant Recovery in order to restore backed up VM's in the event of D/R testing or catastrophe at source site

Am I missing anything?

Am I right in thinking that I need to create a new backup proxy at the source site for use with the Veeam install at the D/R site? Would this be
termed as a source proxy?
Create a proxy at the D/R site which would be the destination proxy?

How much storage approximately would I need at D/R Site to accomplish this?

Thanks in advance

Simon

[foggy]

Hello, Simon. Please review this topic for possible approaches to offsite replication of backup files.

If you decide to go with two jobs, then a couple of thoughts below:

Setup the backup schedule at D/R site not to coincide with that at the source

Scheduling conflicts are handled automatically in Veeam B&R.

Am I right in thinking that I need to create a new backup proxy at the source site for use with the Veeam install at the D/R site? Would this be
termed as a source proxy? Create a proxy at the D/R site which would be the destination proxy?

You can use the default proxy installed as a part of Veeam B&R at the source site with the limitation that the number of concurrent tasks per proxy are counted within each backup server independently. For the destination proxy, the default proxy in Veeam B&R in target site can be used.

How much storage approximately would I need at D/R Site to accomplish this?

You can calculate the required space based on the number of restore points you are required to keep in DR location and the size of your current backup files. You need some space to power up your VMs via Instant Recovery also.

[jumbo22]

Thanks Foggy.

The text listed under 'I Will', would that work? (taking into consideration your earlier response ref proxy's)

Cheers

Simon

[foggy]

Yes, looks like a plan.

[jumbo22]

Thanks Foggy

[AdamAlt]

Hi

I have a Veeam server running 8 restore points, Forward incremental, Full backup on a Friday, Synthetic full on a Saturday.

I want to move a full backup file off site once a week (saturday or sunday) to our offsite backup provider (http://www.backupdirect.onlinebackup.com if your intersted, highly reccommended), so in the event of a restore i can use the local copies of up to a week, but in the event of a total disaster (ie. lose local copies) then i can recover a full backup from a week ago (cannot rely on any incrementals being available)

I'm trying to work out the best way of doing this.

Problem is,I have to select a folder and all of its contents in the Online backup, so the "Backup" folder would be too much as i would be uploading all the incrementals which would cost too much to upload.

Possible solutions

1. Create a batch file that movies a *.vbk to a different folder that i have selected to backup all file contents of to Online backup.
2. Create a seperate backup job "Backup 2" that runs just a full backup once a week and back this up Online. Changing current job to just do incremental & synthetic full, no full backups. (Would that work??)

Any thoughts / help would be hugely appreciated, i've been stuck listening to hold music for 30mins now at my companies expense!!

Thanks

[foggy]

Please review this thread for possible ways of replicating backup files offsite.

Also, I would note, that running both synthetic and active fulls in a row in one job does not make much sense as active full resets the backup chain and starts it anew (keeping previous restore points according to retention settings). So your Saturday synthetic full will be built based on the Friday's active full and Saturday's increment only. Best practice is to run weekly synthetic fulls and run active full backup at least once a month.

[BOK]

Hi,

I've used Veeam for a while, and now I've got a challange where the local backup stored on a NAS box are going to be replicated to a remote site. On both sites they have a QNAP NAS with rsync setup.

So, I wonder which backup strategy would be the best when having bandwith in mind, approx 20 mbit/s and the first backup is about 600GB...?

We want/need a strategy with the bullets below:

1. Daily backup
2. Monthly backup
3. Annualy backup

Thanks!

/BOK

[rubeng]

Hi,

Our case: We have 2 datacenters where I work. The two datacenters is in two different buildings, connected to each other by fibre.

Datacenter 1 is where all our customers is (our main datacenter), and there's also a backupserver there running Veeam Backup & Replication 6.1 (we call that VMbackup1).
In Datacenter 2, we have our internal servers, and two backup servers (VMbackup2 and VMbackup3).

VMbackup1 is taking backup of our customers servers, and VMbackup2 is taking backup of our internal servers.
VMbackup3 is running a synchronization software, called Super-Flexible. Super-Flexible is synchronizing the Veeam backup from VMbackup1 & 2, to VMbackup3. So, all Veeam backup from Datacenter 1, is also located on a backup server in datacenter 2.


Whats the best practice when it comes to backing up your Veeam backup?`
Is there any better solution to this, than the one we already have?



All tips and help will be highly appreciated.

[dellock6]

Hi Ruben,
your design sounds solid. Since Veeam can only use one repository per backup job, we all rely on external tools to replicate backup data somewhere else. So you design isas good as other using san replication or rsync or other tools.
One doubt only: customers backups are saved in dc1 and then in dc2, while your backups are saved twice in the same datacenter; a possible enhancement could be to replicate also vm-backup2 in dc1.

Luca.

[rubeng]

Hi Ruben,
your design sounds solid. Since Veeam can only use one repository per backup job, we all rely on external tools to replicate backup data somewhere else. So you design isas good as other using san replication or rsync or other tools.
One doubt only: customers backups are saved in dc1 and then in dc2, while your backups are saved twice in the same datacenter; a possible enhancement could be to replicate also vm-backup2 in dc1.

Luca.

Yes, I agree.

Thank you for answering me. After doing some more research on this scenario, it seems that there's a lot of people facing the same problems.


Hopefully there will be a feature within Veeam that backups to several repositories!


Have a good weekend:-)

[dellock6]

I don't remember where and when I wrote here about this, the idea could be to have a new kind of repository that can replicate itself, even selectively, or even better a new role maybe called "replicator" whose task is to take data from a repository and copy them to another one. This would be really awesome

Luca.

[EnriquePhoenix]

Well said dellock6. While this can be done easily manually, if it were part Veeam life would be much simpler

[jmacdaddy]

Instead of replicating repositories, does anyone here just make regular backups of their Veeam replicas at their DR sites for long term backup archival?

[foggy]

Sure. But doing so you should note, that VMware CBT cannot be used for replicated VMs.

[herman.rutten]

Hi everyone,

I'm currently designing a backup environment where I have to keep onsite backup copies as well as ofsite backup copies. The most important challenge is the WAN link in-between sites. No tape is involved; all backups should be to disk. The VMware environment is currently being replicated on the storage level.

A thought crossed my mind to stop using the storage replication and opt for Veeam 6.5 replication instead. The strategy would be to create local backup jobs for the original VM's and to create remote backup jobs for the replica VM's at the remote site. This would allow for replication traffic to traverse the WAN only once. However, I'm not sure if Veeam support backups of replica VM's, since I've not been able to find it in any documentation.

Please could anyone elaborate on this?
Many thanks in advance!

Kind Regards,

Herman.

[veremin]

However, I'm not sure if Veeam support backups of replica VM's, since I've not been able to find it in any documentation.

This scenario should be fully supported, as many customers so far have been reporting backing up successfully their replicas without any issues.

Though, be aware that in this case backup of replicas will not be able to use CBT, due to the fact that CBT can’t be enabled on VM with a snapshots, and replica always has at least one. So, it’s likely to be some performance impact.

In addition, little piece of information regarding CBT - it has been also reported several times, that CBT appears not to work with powered-off VMs. Keep it in mind,as well.

Thanks.

[lsittech]

Hello All,

I am new to Veeam however am so great full to be rid of tape backups. Especially the part where I have to take them home every day.

I know this topic has been discussed a lot, I read most of it, just a little confused with it all now.

We currently have 2 x physical hosts clustered with 9 VM Guests across them.
We use Veeam 6.5 to backup these to a local 20TB SAN for easy restore.

I now need to start getting these offsite again. We have storage in a data center available to us. I just need to know the easiest supported way to get them there.

So many people have different suggestions however this is hopefully just a nice easy one.

Should I run second jobs which backup to offsite or simply each night copy the backups sitting on the SAN over to the data center? Should I do so within Veeam or people have suggested use DeltaCopy

Any advice would be greatly appreciated.

Cheers,
Stephen

[jpeake]

secondary job, pointing to the remote storage (install repo using agent). Pre-seed the remote repo if possible, then just run forward incs with sythentic fulls. Link the offsite job to start when the onsite job finishes. Done.

There are a million ways to skin this cat. This method works well for me.

[Vitaliy S.]

Hi Stephen, I would also suggest not to bother with scripts (3rd party tools) and separate retention policies, secondary backup job sounds like the simplest way to achieve what you want. Thanks!