I run into one problem over and over again ...
I am inside an ESXi environment that I have never seen before and need to upload or download a large amount of data - say 1TB or so - from an ESXi host.i
Troubleshooting performance issues in that environment is not an option as I am in a hurry.
So I need to pick one option that has good performance and a very high success rate - with one TB of data to copy in a friday night failures are not an option.If a procedure tends to failures I need the option to resume from where it failed last time.
Scenario for an example: a VMFS 5 datastore with a 1Tb flat.vmdk - must be downloaded - and uploaded.
here are some possible procedures: is there one that should be faster than others ?
- copy via Datastorebrowser to a Windows host = slow and unreliable
- copy via SCP (WinSCP or something like that)
- copy via FTP (Veeam FastSCP for example)
- vmkfstools -i to a remote NFS-server
- dcfldd to a network share (slow - but can be resumed after failures)
- ghost32 into another vmdk
- robocopy into another vmdk
... just to name a few options ..
What options should be fastest ?
What is faster - using a local host connected to an ESXi - or using a local host connected to a VM inside that ESXi ?