New release will re-pull the entire image.
It was working in the past, pulling only new layers on the top. But now it is pulling everything again. Slow to do testing.
10 Replies
Are you using the same tags?
Also not sure if there's some update on Runpod's side
Interesting. I am using the same tag but only increase in version number.
we updated it because the previous implementation had a bug, we plan to introduce that back eventually
Gotcha. Thanks. I suppose it will waste a lot of traffic...
This slows my testing much. When will this be introduced back?
ill provide an update early next week, already have pushed this in engineering to put a fix out and introduce the optimization
If i may suggest, it's faster if you just use pods as a dev environment it can be faster
plan to release the optimization this upcoming week, testing in dev
any update here?
it should be pulling only new layers, when release occurs, we update workers on the same servers to better utilize cache, if you see something different share more details
I keep seeing this for out endpoint with 31 GB images. Here are the details of a recent release:
Endpoint: rcq6j9fatqb4n5
Vold: 35-51db7ee
Vnew: 36-2338b79
I see that only several top layers has changed, yet it downloads gigabytes. Could you please take a look?