Read the entire restore point from source backup

[andriktr]

Hi all,
In our environment we have few backup copy jobs which are copying backups for long term storing with GFS retention to another Site/Data Center. These copied backups stored on the HP Store Once.
After upgrading to Veeam v9 I discovered new option in job configuration "Read the entire restore point from source backup instead of synthesizing it from increments". After enabling this option backup copy jobs runs few times faster than before.
Maybe someone can explain in little bit more details about differences when I run job with option and without.
Thank you.

[foggy]

It does just that - reads the entire full backup data from the source repository instead of synthesizing it from files stored on the target one (which requires random I/O + data rehydration).

[andriktr]

So if I understand correctly in my case it's recommended to turn on this option. Yes?

[foggy]

Correct.

[andriktr]

Thanks for help

[aarumugam]

Hi Foggy,
I have a follow up question, to your below response.

It does just that - reads the entire full backup data from the source repository instead of synthesizing it from files stored on the target one (which requires random I/O + data re-hydration).

Just to make sure, I understood it correctly, synthesizing from files stored on the target one (which requires random I/O + data re-hydration) inst it same as Doing a Synthetic Full Transformation ?
Also, reading from the source will also do the same right, creating a synthetic full which also requires random I/O + data re-hydration

Is one of the prime reason for enabling "reads the entire full backup data from the source repository instead of synthesizing", Is because reading from source would be faster, as they have a shorter backup chain, compared to a GFS which will mostly have a longer backup Chain

As per my understanding both do the same operation of synthetic full, only from which repository is the question. Either Source or Target Repository.

BTW I did read the link you had in your post, that still did not answer this.

[foggy]

If the "Read from source..." option is enabled, entire VM data is read from the source repository and is sequentially written to a brand new full backup file, no synthetic activity is performed (so this is completely similar to active full for regular backup jobs). If the option is not enabled, the oldest increment will be merged into the existing full backup during each job cycle (similar to forever forward incremental mode).