According to the docs, memcached
is used by flux
to store image metadata. Fair enough. I tried connecting to it and running some commands, and I was surprised to see that flux
actually pulls all images from a all registries and keeps them in memcached
. Why is it doing it?
For example if you check calico/cni image, there are 103 pages (25 images per page) of images. So, that's more then 2500 image metadata to store in memcached
. Is this necessary? Am I ever going to need images 2-3 years old?