Fastest Method for Internal File Transfer Speed between Disk Array and VM/Dockers

[rvijay007]

Hi all,

 

Just wanted to ask your collective wisdom on what the best method for mounting an unRAID user share within a VM was, for read file access. I've seen several threads on different methods across the years. I am in an Ubuntu environment. So far, I've tried:

 

• Mount via the unRAID mount tag specified under VM settings and access/copy file
  • sudo mount -t 9p -o trans=virtio iso_files /mnt/local_isos/ ; pv /mnt/local_isos/ubuntu-20.04-desktop-amd64.iso > ubuntu-desktop.iso  
    • Speed: [28.6MiB/s]
• Connect to share via Samba and access/copy file
  • pv '/run/user/1000/gvfs/smb-share:server=tower,share=isos/ubuntu-20.04-desktop-amd64.iso' > ubuntu-desktop.iso

    • Speed: [ 252MiB/s]

• Copy from one vdisk to another vdisk attached to same VM

  • pv ./ubuntu-desktop.iso > ../LargeSwap/ubuntu-desktop.iso

    • Speed: [1.02GiB/s] (view this as upper limit)

 

Samba access seems to be 10X faster using the pv command to do a copy than the 9p internal mount method, but I'm wondering if I'm missing anything as I expected 9p to be equivalent? 

 

Also, is there an even faster method that I'm not aware of for doing these internal file transfers? Does Samba, even on a local VM connecting to the same on computer disk array, congest the network by hitting the router with the file transfer?

 

Thanks so much for your help!

 

 

P.S. I'm a bit of an Ubuntu newbie, so if you have any new methods that are faster, details on how to set up would be much appreciated!

Edited July 14, 2020 by rvijay007

[rvijay007]

bump