Hi dear community,
We are hosting private docker registry on different remote network and experiencing some pull issues.Now overall scenario is imagine 3 different remote networks.1 and 2 networks are clients and locate on same country 3rd network is docker registry and locates on different country that clients.Now 1st client has like 936.75Mbps(Latency: 1.47 ms (0.12 ms jitter)) download speed and almost has no issues pulling docker images.But second client network has a bit slower (generally high) internet speed like 100Mbps(Latency: 1.44 ms (0.55 ms jitter)).Now this second client has pulling issues like sometimes pull starts fast on some layers then one of the layers(this layer always changes ,not same layers) starts pulling very-very slowly like Kilobayts per second and that pulls takes a lot of time to pull.If I am lucky docker will retry pulling of layer and right after this retry layer pulls fastly or I would need to restart docker service, restarting docker service helps for that particular time(but very bad idea to restart docker on production).I have 3 questions
1)What can be the reason for this problem? is there any way to full debug pull process?
2)How can I customize retry interval for docker layers?
3Is there any way to force docker retry image pull without restarting docker(image pulls are done by kubernetes so pull works somewhere in background)