Recently I had to move some VPSes from one server to another, as the old server was being shut down (new cheaper and faster host). To minimize the downtime, I took data from one LVM volume and put it directly in the LVM of the new server.

Receiving side:

time nc -l 19000 | pigz -d -k -c - | dd of=/dev/vps-vg/vps04

Explanation:

• I use “time” just for stats (so it's not required)
• I set up a listening netcat socket on port 19000, awaiting data from the “sending side”
• I decompress incoming data from netcat with pigz (“-” is for data coming from “stdin”)
• Finally I put the decompressed data into the new LVM volume with “dd” (dd grabs its data from stdin if no “if” parameter is set)

Sending side:

dd if=/dev/vg0/vps04 bs=4096k | pigz -9 -b 4096 -c | nc RECEIVING.SERVER.TLD 19000

Explanation:

• I read the data with dd
• I compress it with pigz
• I then use netcat to send the data.

NOTE: You need to start the “receiving side” first, otherwise the “sending side” will fail due to netcat being unable to establish a connection.