I test the dragonfly for image distribution efficiency. I have 2 supernode, 18 dfclient. I simulate dfclient pull 20G image at some time with crontab. Only 2 dfclient can pull successfully. Others occur Get https://registry-1.docker.io/v2/: dial tcp 34.232.31.24:443: connect: network is unreachable error, some image mirror layer print output like 60b1ddcd47df: Retrying in 1 second. I restart supernode, pull again at pull error machine, it can pull succeccful. Why?
Q:
Have you test really the image distribution efficiency for simulating pull request at same time with installing crontab work?
What you use for docker image distribution at your department?
Question
I test the dragonfly for image distribution efficiency. I have 2 supernode, 18 dfclient. I simulate dfclient pull 20G image at some time with
crontab
. Only 2 dfclient can pull successfully. Others occurGet https://registry-1.docker.io/v2/: dial tcp 34.232.31.24:443: connect: network is unreachable
error, some image mirror layer print output like60b1ddcd47df: Retrying in 1 second
. I restart supernode, pull again at pull error machine, it can pull succeccful. Why? Q: Have you test really the image distribution efficiency for simulating pull request at same time with installing crontab work?What you use for docker image distribution at your department?
config:
a part info for
docker info
:supernode start cmd:
dfclient start cmd: