NAS backup running for a very long time

Hi ​@ryan.drm ,
we had a similar issue with one of our customers. What helped in our case was adding more proxies and reviewing the network setup.

In this specific situation, the customer was in the process of migrating from their old data center to ours. As a result, the traffic was routed across multiple firewalls and effectively across the entire country before reaching the backup infrastructure.

We also removed the backup server from the proxy list because it was located in a third data center, which added unnecessary complexity.

To optimize network traffic, we added a second NIC to the proxies, which eliminated the routing issue until all networks and routing could be consolidated within our data center.

Additionally, we restructured the backup jobs by using exclusions and creating dedicated jobs for large folders within the share. This allowed us to back up smaller chunks, which significantly improved performance.

With all these adjustments, we were able to increase the throughput to around 800 MB/s and reduce the overall backup duration. The initial backup of very large folders still took some time, but the subsequent runs were within an acceptable time-frame for the customer.

The next step will be to deploy additional proxies and remove the second NIC once the migration to our data center is fully completed.

Regards

Chalid

Hi Chalid,

Thank you for your response and explanation. 

Currently, all components (VBR, proxy, repository) are located in one area and use a 10GBps network. 

For the backup job that is currently running, can we move it to the new proxy, or do we have to cancel it first and then move it to the new proxy? Assuming that the new proxy is a VM located on Nutanix AHV. 

And does the repository cache need to be increased as well? If so, can we use the Hardened Linux Repository as a cache repository in addition to the VBR server itself?

And is the method used “backup directly from file share” or “backup from storage snapshot”?
Currently, we are performing backups from the Nutanix file share. 

Hi ​@ryan.drm 

unstructured data backup are mostly different as VM backup. Now you are only at 10-15% after 140 hours, this means several more weeks to finish. If you cannot tolerate that, it is likely better to cancel the job, re-architect the proxies/method (see below), and restart. 

As first step I think best way is to backup one share at time due to their biggest size.

You must implement multiple General Purpose Backup Proxies and a Cache repository where store metadata. You can use one of GP Proxies but I recommend to use SSD disk. 

Alternatively you can use a storage snapshot backup adding a Nutanix File server in Veeam infrastructure considering all requisites and limitation for your environment. 

My 2 cents