Veeam and CAtalyst scenario

[Andanet]

Hi all,
this question is to verify other experiences using HPE Catalyst store on Storeonce.
Now our environment has 1 Veeam server, two phisical proxy and 3 store catalyst.
backup job use both proxy but there are written in 2 catstore which have one or other proxy as gateway to eliminate extra hop and reduce CPU Utilization.
Third catstore is used for a copy job with GFS retention and has not gateway setting.
I have verified that store with GFS has a better deduplication than others. Logically there are more data and dedupe ratio is better.
I'm thinking to move backup job on one catstore (where copy jobs writing) but I've a doubt and a question for all:
Gateway configuration on this catstore (proxy affinity is automatic) must be modified or can I leave default?
What kind of impacts will be there for backup job and its proxy? Hops from proxies will be incremented?
Thanks for your attention.
Antonio

[foggy]

Hi Antonio, do you mean that currently you have several backup jobs going to different Catalyst stores and select the proxy in each of them explicitly, depending on the repository it is pointed to? If you leave explicit proxy selection, then everything should be fine, even though the gateway is selected automatically.

In case of automatic proxy selection though, there's a probability that some of the VM disks will still be processed by the "wrong" proxy, so enabling per-VM backup chains for repository is recommended to eliminate such cases. Please review this thread for some additional details.

[Andanet]

Thanks foggy.
Proxy in job are automatic selected. But store where restore point has saved have gateway set with one of its.
Per VM chain is by default and correctly used.
I think new solution is pointing all backup job to on e catstore using both proxy in job configuration and using affinity for the store without set gateway server.
Is possible to do this? What kind of network traffic throught proxies we obtain in this case?
Another catstore can be used for copy job with GFS retention.
Thanks for link to other discussion.

[foggy]

Yes, such setup is even more effective in terms of excessive network traffic, since it will result in less cases of "wrong" proxy selection (less traffic between proxies). Setting proxy affinity is not required here, since both proxies are allowed to write to this repository.