The replication speed depends on 4 key factors:
- The uplink speed from that server to the replication server and bandwidth available.
- The overall disk storage.
- The changes in the disk while it is replicating.
- I/O speed of the storage itself.
To test your uplink speed, you could iperf3 utility, for example, as follows:
- Install a vanilla Linux machine (m4.xlarge on AWS, Basic_A2 on Azure or n1-standard-1 on GCP) in the same subnet of the CloudEndure replication servers
- On that machine, please install iperf3 utility using:
sudo apt-get install iperf3
**If you're using Red Hat/CentOS use yum install instead. Please note that epel-release package may be required. In that case, see 'Method 2' here.
- Then run:
iperf3 -s -p 1500
- On your source server, please install iperf3 as well
- Windows: Download the right zip file from here and extract it.
- Linux: Install as mentioned above.
- Then, on terminal window please run:
iperf3.exe -i 10 -c [linux server ip] -p 1500 -t 60 -P 1
Here is a sample output:
In this output, you can see that the uplink is 23.4Mbps which means that a 100GB (idle) server should be replicated in about 10 hours. You can use this calculator, for example.
Please note, if your server writes to disk an average of 20GB/day you need to take it into account when calculating as follows:
20GB/day --> ~2Mbps --> which leave us with only 21.4Mbps available for the initial 100GB.