What resources are used during Transform?

[readie]

Our transforms are taking a very long time, and I would like to try to ensure that resources are not being used which could slow down our production system.
I appreciate that the storage is in our backup data cabinet (on site, but a distance from production), but is the main Veeam server involved in reading and writing data, or is it one of the proxy servers (I have selected 'automatically choose Proxy' for that job) so might it be any of them (we have two separate in addition to the Veeam server). I presume that will be network traffic, or will the Proxy also be using its HD (over iSCSI) which might impact our production iSCSI?
I would appreciate a clear explanation/block diagram - or a reference to any already existing - so that I can plan our infrastructure suitably?
thanks,
Bob

[dellock6]

I Bob, I don't have a diagram at hand, but the transform operation is processed by the repository.
The only reason the veeam server is involved, is if IT is also the repository, as for example when using a file share on CIFS as a repository.

If you have a dedicated machine acting as a repository, that one is running transforms. The related section in the user guide:
http://helpcenter.veeam.com/backup/70/v ... 1363540327
explains that the job involves a repository, and the transform counts as an active task, so be careful when configuring task limits in the repository.

Luca.

[readie]

Ok thanks. We are using a CIFS fileshare . . . so if instead we use a Windows server (attached to the same CIFS share) and define it as the repository, then it will do the transform without using the Veeam server? That would be better, particularly as we can have the windows server physically and logically close to the fileshare, so that it doesn't impact production. I will have a look at the manual too.
Bob

[veremin]

Hi, Bob, in case of CIFS repository transformation activity is performed on proxying server. By default, the role of proxying server is assigned to VB&R management server. If you want to change it, go to the setting of the repository and specify there whatever server you want to.

Thanks.

[readie]

Thanks - I'm happy to use any Proxy for the actual overnight backups (we have three sharing the jobs) - it's just the transform that I would like to allocate to a specific machine. So am I right in thinking that if I make the repository a Windows server (attached to the CIFS share) then I can continue to use my three proxies (all sending backups to the Windows server) but the transform will be performed by the Windows server, not the proxies?
(and a supplementary question, if I change the job from CIFS to Windows 2012 server pointing to the same CIFS share, will it 'see and know about' the existing backup seeds, or will I have to start again with another full backup? The backup system is on-site, using LAN not WAN.)
Bob

[dellock6]

Hi,
indeed you can dedicate a server to act as a repository, and it will run any related activity.
Only be aware you cannot map a CIFS share in direct attach to a server, share mounts are in user space, cannot be used by the system itself.

About mapping, yes you can for sure re-use the backups and start from there. Once you added the new repository, rescan it in order to import backups, and inside the involved jobs use the "map backup" feature to map the backup job to the existing backups.

Luca.

[veremin]

Thanks - I'm happy to use any Proxy for the actual overnight backups (we have three sharing the jobs) - it's just the transform that I would like to allocate to a specific machine. So am I right in thinking that if I make the repository a Windows server (attached to the CIFS share) then I can continue to use my three proxies (all sending backups to the Windows server) but the transform will be performed by the Windows server, not the proxies?

If you attach CIFS share to Windows based machine, and add this machine as Windows repository, then, the transformation activity will be isolated to this machine, and not other machines will be involved.

If I change the job from CIFS to Windows 2012 server pointing to the same CIFS share, will it 'see and know about' the existing backup seeds, or will I have to start again with another full backup?

In order to do so you should use backup mapping functionality.

Thanks.