Presenting a large volume to virtualised veeam

[malds1]

Hello,

We have have a large 30TB direct attached ESX host in a secondary site which we plan to use as a DR/Backup solution - not ideal but fulfils the purpose of backups and continently of core services.

The plan is to have Veeam take daily / monthly backups that will be written to a virtualised repository on that host (which will then go onto tape) and to have Veeam continuesly replicate the mission critical services to the same ESX host & datastore.

I have two questions - what would be the best way to present the local 30TB of storage to the Virtualised Veeam instance on the DR host for backups, the plan being to allocate it 15TB - the other 15 being used for replication jobs.
- VMDK's and span the volumes doesn't strike me as the most efficient method.
- the other is that we are getting lots of very spikey transfers when doing network backups, the source shows a constant access rate to the disk whereas the network on the Proxy shows a very spikey 3-40% continues short interval spikes, is this as a result of the compression, the repository then also has these spikes.

Greatly appreciate your your views on the above,

T

[Gostev]

Hi,

1. You should use RDM (raw device mapping) disk in physical mode.
2. Yes, both compression and source-side dedupe (some duplicate blocks are rejected after being read).

Thank you.