-
- Enthusiast
- Posts: 64
- Liked: 10 times
- Joined: May 15, 2014 3:29 pm
- Full Name: Peter Yasuda
- Contact:
Architecture and data flow questions
I have a single server with all the Veeam B&R roles. It has a 3 gigabit NICs, one to the LAN, and two to an Equalogic SAN. Storage is a SAS disk array.
Most of the backups are direct from the SAN, using the server as an off-host proxy, and some backups are from off-SAN servers through VAW to an SMB share.
I am considering adding storage by purchasing a new storage server, but only if the single gigabit LAN connection won't be a problem. In this scenario, the original server stays as is, and the new server has the Repository role.
If I move the VAW backups to a repository on the new server, will backup data go directly to that server, or relay through the backup server?
Does it matter whether I add a proxy role to the repository server?
Will the work done to merger forever forward incrementals be done on the Repository, without needing a lot of data to go back and forth between the backup server?
Another option is to move all the roles to a new server, and use the current server as just a second repository, but that would be a higher level of effort, and require loading a new server with Server 2012 (for optimal compatibility with the SAN and Hyper-V cluster).
Most of the backups are direct from the SAN, using the server as an off-host proxy, and some backups are from off-SAN servers through VAW to an SMB share.
I am considering adding storage by purchasing a new storage server, but only if the single gigabit LAN connection won't be a problem. In this scenario, the original server stays as is, and the new server has the Repository role.
If I move the VAW backups to a repository on the new server, will backup data go directly to that server, or relay through the backup server?
Does it matter whether I add a proxy role to the repository server?
Will the work done to merger forever forward incrementals be done on the Repository, without needing a lot of data to go back and forth between the backup server?
Another option is to move all the roles to a new server, and use the current server as just a second repository, but that would be a higher level of effort, and require loading a new server with Server 2012 (for optimal compatibility with the SAN and Hyper-V cluster).
-
- Product Manager
- Posts: 8191
- Liked: 1322 times
- Joined: Feb 08, 2013 3:08 pm
- Full Name: Mike Resseler
- Location: Belgium
- Contact:
Re: Architecture and data flow questions
Hi Peter,
I think the biggest question for you here is: Can I "avoid" that the data goes to the backup server. If I understood that correctly, then the answer is yes. The backup server will still has some traffic as it is the 'control-and-command' center so those still need to be sent (otherwise not much will happen ). But the data itself is moved between data movers. So as long as you make sure that your backups are being sent to the right datamovers (proxies) it won't pass the backup server.
If you (quickly) read this process here: https://helpcenter.veeam.com/docs/backu ... tml?ver=95 you will be able to see the different steps when a backup is taken.
Obviously, once you get started with this, you will need to configure your jobs so they use the correct proxies.
Makes sense?
Mike
I think the biggest question for you here is: Can I "avoid" that the data goes to the backup server. If I understood that correctly, then the answer is yes. The backup server will still has some traffic as it is the 'control-and-command' center so those still need to be sent (otherwise not much will happen ). But the data itself is moved between data movers. So as long as you make sure that your backups are being sent to the right datamovers (proxies) it won't pass the backup server.
If you (quickly) read this process here: https://helpcenter.veeam.com/docs/backu ... tml?ver=95 you will be able to see the different steps when a backup is taken.
Obviously, once you get started with this, you will need to configure your jobs so they use the correct proxies.
Makes sense?
Mike
-
- Expert
- Posts: 109
- Liked: 5 times
- Joined: Jul 13, 2017 12:34 pm
- Contact:
Re: Architecture and data flow questions
I agree with Mike, that you don't need to send the Data to the Backup Server.
I don't agree about the Proxy, as in this case it's not called Proxy but Gateway Server.
You need to add the new Storage Server as a Managed Server in B&R and then choose this Server as the Gateway/Location of your new Repository.
It doesn't need the Proxy role, as it is not connecting to the to-be-backed-up Servers via SAN and doesn't copy the VSS's directly from it, rather it receives the Data from the On-Host Proxy/the VAW and then saves it to the Repository.
So basically like this: Hyper-V host as On-Host Proxy -> LAN -> Gateway Server/New Repository
I often see that Gateway Servers are called Proxies, but that isn't the same and got me into a lot of trouble in my environment (the backup job will fail if you set any off-host proxy without SAN connection and the right Hardware VSS's)
I don't agree about the Proxy, as in this case it's not called Proxy but Gateway Server.
You need to add the new Storage Server as a Managed Server in B&R and then choose this Server as the Gateway/Location of your new Repository.
It doesn't need the Proxy role, as it is not connecting to the to-be-backed-up Servers via SAN and doesn't copy the VSS's directly from it, rather it receives the Data from the On-Host Proxy/the VAW and then saves it to the Repository.
So basically like this: Hyper-V host as On-Host Proxy -> LAN -> Gateway Server/New Repository
I often see that Gateway Servers are called Proxies, but that isn't the same and got me into a lot of trouble in my environment (the backup job will fail if you set any off-host proxy without SAN connection and the right Hardware VSS's)
-
- Enthusiast
- Posts: 64
- Liked: 10 times
- Joined: May 15, 2014 3:29 pm
- Full Name: Peter Yasuda
- Contact:
Re: Architecture and data flow questions
Thank you for your answers, and the link to the backup steps. I knew I had seen it before, somewhere!
I think in my case I will not need the gateway server, because the storage will be directly attached to the repository server. SMB will not be needed because the VAW backups can be sent directly to the repository. So I will not be in the 3 situations below:
A gateway server is an auxiliary backup infrastructure component that “bridges” the backup server and backup repository. The gateway server is required if you deploy the following types of backup repositories in the backup infrastructure:
Shared folder backup repositories
Dell EMC Data Domain deduplicating storage appliance
HPE StoreOnce deduplicating storage appliance
I have used a gateway in other situations where the backup target was a SMB NAS.
I think in my case I will not need the gateway server, because the storage will be directly attached to the repository server. SMB will not be needed because the VAW backups can be sent directly to the repository. So I will not be in the 3 situations below:
A gateway server is an auxiliary backup infrastructure component that “bridges” the backup server and backup repository. The gateway server is required if you deploy the following types of backup repositories in the backup infrastructure:
Shared folder backup repositories
Dell EMC Data Domain deduplicating storage appliance
HPE StoreOnce deduplicating storage appliance
I have used a gateway in other situations where the backup target was a SMB NAS.
-
- Expert
- Posts: 109
- Liked: 5 times
- Joined: Jul 13, 2017 12:34 pm
- Contact:
Re: Architecture and data flow questions
"because the storage will be directly attached to the repository server"
That means your "repository server" will take over the role of the Gateway Server, which a NAS/SMB Share cannot
That means your "repository server" will take over the role of the Gateway Server, which a NAS/SMB Share cannot
Who is online
Users browsing this forum: lando_uk, Semrush [Bot], veremin, whistler and 159 guests