Christian Richter
Christian Richter
I think the issue happens only with corrupt databases, I could not reproduce it for the last weeks....
Just to sum up things here we talked about in the meeting: The idea here is to piggy back on the helm getters (port them over to hypper) and create...
WIP implementation can be found in this fork/branch https://github.com/dragonchaser/helm/tree/add_download_cache
Ok, going for low hanging fruit then and use 2 caches.
HIP for upstream HELM PR is here: https://github.com/helm/community/pull/185
@mattfarina I am having a bit of a hard time to track that down, can you give me a more detailed workflow example how this happens? Thx.
Because it relies on the Install methods of helm and if it is installable is not determined on our side atm, meaning: if the chart is not resolvable it will...
> imho, an optional dependency should be similar to a [zypper recommends](https://en.opensuse.org/Libzypp/Dependencies#Weak_dependencies) which "are silently dropped if no match is possible." Ah ok now I understand, still we need to...
This should also feature detecting stale deployments, e.g. dependencies that have been installed to the cluster but are not used by any "parent" deployment any-more.
possible duplicate of #42