It has a few implications to launch a raft peer when you wanted to do crdt and vice-versa.
Same when exporting and exporting states.
Users starting cluster peers should be explicit about their consensus choice.
Also, if we ever want to make `crdt` the default, we can't do that before
making `raft` non-default first. I don't like to break things but otherwise
the experience for new users wanting to try crdts might be aweful.
By using gx install --local, deps are copied into
the build context and therefore only stale deps
will get pulled from the network on image build.
License: MIT
Signed-off-by: Adrian Lanzafame <adrianlanzafame92@gmail.com>
The main container will now run only ipfs-cluster-service.
A new ipfs-cluster-bundle container is built by Dockerfile-bundle
which will provide ipfs-cluster+ipfs.
Fixes#197
License: MIT
Signed-off-by: Hector Sanjuan <code@hector.link>
This used to work. Even if it's been surfaced by 4a8759939 we were
extensively testing the Dockerfile before that release.
License: MIT
Signed-off-by: Hector Sanjuan <hector@protocol.ai>
After changes on how the final docker image for go-ipfs is generated,
our Dockerfiles failed to build.
The new version is inspired on go-ipfs Dockerfile. It takes advantage
of docker >= v17.06 which allows multi-stage Dockerfiles with
several FROM directives and COPY --from. This allows build freely on a
first stage, without thinking of cleaning up, and then just copy the
final artifacts to a second, minimal image.
The Dockerfile-test is similarly updated. Minor changes have been
applied to the entrypoints.
License: MIT
Signed-off-by: Hector Sanjuan <hector@protocol.ai>
Updates dockerfile and entrypoint script to have the same behavior the ipfs-go docker build
moved to in 3477. This fixes a bug where the cluster entrypoint script does not have
sufficient permissions to access /data/ipfs. The entrypoint also now launches a process with
tini