install application dependencies during the image build #335
Conversation
| @@ -1,5 +1,9 @@ | |||
| FROM node:8.9 AS build | |||
| WORKDIR /app | |||
|
|
|||
| COPY package.json yarn.lock /app/ | |||
| RUN yarn install --frozen-lockfile | |||
There was a problem hiding this comment.
This improves docker cache hits in the development workflow
There was a problem hiding this comment.
what do you mean with "this"? the --frozen-lockfile? Why not doing the yarn install after the COPY . /app?
There was a problem hiding this comment.
the --frozen-lockfile?
Apparently yarn install could go an update the yarn.lock file and update the package versions to something completely different from what it was tested against. For this reason the --frozen-lockfile is recommended for use in CI builds yarnpkg/yarn#749
Why not doing the yarn install after the COPY . /app?
Specifically w.r.t the developer use-case, this ensures that the docker layer cache gets reused during the builds. As a result image builds will be quicker when the dashboard sources (css/js/ts/etc) are updated. Only when there are updates to the package.json/yarn.lock file, there will be a cache miss and yarn install will execute again.
sameersbn
force-pushed
the
fix-dashboard-dockerfile
branch
from
3495367 to
237da38
Compare
| @@ -1,5 +1,9 @@ | |||
| FROM node:8.9 AS build | |||
| WORKDIR /app | |||
|
|
|||
| COPY package.json yarn.lock /app/ | |||
| RUN yarn install --frozen-lockfile | |||
fixes #332