In need to ask you some questions - where does your repo reside at and which machine serves as a proxy?Would I be able to backup several VMs at the same time using all the 4 NICs on the Veeam server considering the Veeam host and all hosts where the VMs are using 4 NICs?
PTide wrote:Then I suggest you to add all four pNICs of the ESXi host where your Veeam server resides at to the same vSwitch and configure NIC teaming. After that, create "Backup Network" on that vSwitch and add "Backup Network" vNIC to Veeam server. That will allow your Veeam repository to utilize all four pNICs. Next, since your proxies are going to be virtual that allows you to leverage hot-add mode, thus the traffic will go from proxy to backup repository via proxy-VM's vNIC, not via vKernel as it would in case of NBD mode. So yes, you can get 1Gb per VM on the repository side. Don't forget that proxies have to be in the same network with your Veeam server. Please note, that network throughput is not the only thing you should care about - please refer to requirements section.
Vitaliy S. wrote:Before upgrading your network configuration for better job performance rates, can you please tell us what is your current bottleneck statistics?
I believe that you won't be very happy to find out that it's the production storage what causes delays, not repository. I suggest you to deploy test Veeam server and check the bottlenecks statistics, as Vitaliy has already proposed.I'm even thinking about making a 2 tier backup repository, one primary with consumer grade SSDs so it can be written very quickly, then a backup copy to a large RAID5 array of HDDs. What do you think?
mephisto wrote:I'm even thinking about making a 2 tier backup repository, one primary with consumer grade SSDs so it can be written very quickly, then a backup copy to a large RAID5 array of HDDs. What do you think?
Users browsing this forum: calamarz and 21 guests