gristlabs_grist-core/Dockerfile

146 lines
5.1 KiB
Docker
Raw Normal View History

(core) add machinery for self-managed flavor of Grist Summary: Currently, we have two ways that we deliver Grist. One is grist-core, which has simple defaults and is relatively easy for third parties to deploy. The second is our internal build for our SaaS, which is the opposite. For self-managed Grist, a planned paid on-premise version of Grist, I adopt the following approach: * Use the `grist-core` build mechanism, extending it to accept an overlay of extra code if present. * Extra code is supplied in a self-contained `ext` directory, with an `ext/app` directory that is of same structure as core `app` and `stubs/app`. * The `ext` directory also contains information about extra node dependencies needed beyond that of `grist-core`. * The `ext` directory is contained within our monorepo rather than `grist-core` since it may contain material not under the Apache license. Docker builds are achieved in our monorepo by using the `--build-context` functionality to add in `ext` during the regular `grist-core` build: ``` docker buildx build --load -t gristlabs/grist-ee --build-context=ext=../ext . ``` Incremental builds in our monorepo are achieved with the `build_core.sh` helper, like: ``` buildtools/build_core.sh /tmp/self-managed cd /tmp/self-managed yarn start ``` The initial `ext` directory contains material for snapshotting to S3. If you build the docker image as above, and have S3 access, you can do something like: ``` docker run -p 8484:8484 --env GRIST_SESSION_SECRET=a-secret \ --env GRIST_DOCS_S3_BUCKET=grist-docs-test \ --env GRIST_DOCS_S3_PREFIX=self-managed \ -v $HOME/.aws:/root/.aws -it gristlabs/grist-ee ``` This will start a version of Grist that is like `grist-core` but with S3 snapshots enabled. To release this code to `grist-core`, it would just need to move from `ext/app` to `app` within core. I tried a lot of ways of organizing self-managed Grist, and this was what made me happiest. There are a lot of trade-offs, but here is what I was looking for: * Only OSS-code in grist-core. Adding mixed-license material there feels unfair to people already working with the repo. That said, a possible future is to move away from our private monorepo to a public mixed-licence repo, which could have the same relationship with grist-core as the monorepo has. * Minimal differences between self-managed builds and one of our existing builds, ideally hewing as close to grist-core as possible for ease of documentation, debugging, and maintenance. * Ideally, docker builds without copying files around (the new `--build-context` functionality made that possible). * Compatibility with monorepo build. Expressing dependencies of the extra code in `ext` proved tricky to do in a clean way. Yarn/npm fought me every step of the way - everything related to optional dependencies was unsatisfactory in some respect. Yarn2 is flexible but smells like it might be overreach. In the end, organizing to install non-core dependencies one directory up from the main build was a good simple trick that saved my bacon. This diff gets us to the point of building `grist-ee` images conveniently, but there isn't a public repo people can go look at to see its source. This could be generated by taking `grist-core`, adding the `ext` directory to it, and pushing to a distinct repository. I'm not in a hurry to do that, since a PR to that repo would be hard to sync with our monorepo and `grist-core`. Also, we don't have any licensing text ready for the `ext` directory. So leaving that for future work. Test Plan: manual Reviewers: georgegevoian, alexmojaki Reviewed By: georgegevoian, alexmojaki Differential Revision: https://phab.getgrist.com/D3415
2022-05-12 15:24:48 +00:00
################################################################################
## The Grist source can be extended. This is a stub that can be overridden
## from command line, as:
## docker buildx build -t ... --build-context=ext=<path> .
## The code in <path> will then be built along with the rest of Grist.
################################################################################
FROM scratch as ext
################################################################################
## Javascript build stage
################################################################################
(core) move more tests to grist-core Summary: * Tie build and run-time docker base images to a consistent version (buster) * Extend the test login system activated by GRIST_TEST_LOGIN to ease porting tests that currently rely on cognito (many) * Make org resets work in absence of billing endpoints * When in-memory session caches are used, add missing invalidation steps * Pass org information through sign-ups/sign-ins more carefully * For CORS, explicitly trust GRIST_HOST origin when set * Move some fixtures and tests to core, focussing on tests that cover existing failures or are in the set of tests run on deployments * Retain regular `test` target to run the test suite directly, without docker * Add a `test:smoke` target to run a single simple test without `GRIST_TEST_LOGIN` activated * Add a `test:docker` target to run the tests against a grist-core docker image - since tests rely on certain fixture teams/docs, added `TEST_SUPPORT_API_KEY` and `TEST_ADD_SAMPLES` flags to ease porting The tests ported were `nbrowser` tests: `ActionLog.ts` (the first test I tend to port to anything, out of habit), `Fork.ts` (exercises a lot of doc creation paths), `HomeIntro.ts` (a lot of DocMenu exercise), and `DuplicateDocument.ts` (covers a feature known to be failing prior to this diff, the CORS tweak resolves it). Test Plan: Manually tested via `buildtools/build_core.sh`. In follow up, I want to add running the `test:docker` target in grist-core's workflows. In jenkins, only the smoke test is run. There'd be an argument for running all tests, but they include particularly slow tests, and are duplicates of tests already run (in different configuration admittedly), so I'd like to try first just using them in grist-core to gate updates to any packaged version of Grist (the docker image currently). Reviewers: alexmojaki Reviewed By: alexmojaki Subscribers: alexmojaki Differential Revision: https://phab.getgrist.com/D3176
2021-12-10 22:42:54 +00:00
FROM node:14-buster as builder
# Install all node dependencies.
(core) add machinery for self-managed flavor of Grist Summary: Currently, we have two ways that we deliver Grist. One is grist-core, which has simple defaults and is relatively easy for third parties to deploy. The second is our internal build for our SaaS, which is the opposite. For self-managed Grist, a planned paid on-premise version of Grist, I adopt the following approach: * Use the `grist-core` build mechanism, extending it to accept an overlay of extra code if present. * Extra code is supplied in a self-contained `ext` directory, with an `ext/app` directory that is of same structure as core `app` and `stubs/app`. * The `ext` directory also contains information about extra node dependencies needed beyond that of `grist-core`. * The `ext` directory is contained within our monorepo rather than `grist-core` since it may contain material not under the Apache license. Docker builds are achieved in our monorepo by using the `--build-context` functionality to add in `ext` during the regular `grist-core` build: ``` docker buildx build --load -t gristlabs/grist-ee --build-context=ext=../ext . ``` Incremental builds in our monorepo are achieved with the `build_core.sh` helper, like: ``` buildtools/build_core.sh /tmp/self-managed cd /tmp/self-managed yarn start ``` The initial `ext` directory contains material for snapshotting to S3. If you build the docker image as above, and have S3 access, you can do something like: ``` docker run -p 8484:8484 --env GRIST_SESSION_SECRET=a-secret \ --env GRIST_DOCS_S3_BUCKET=grist-docs-test \ --env GRIST_DOCS_S3_PREFIX=self-managed \ -v $HOME/.aws:/root/.aws -it gristlabs/grist-ee ``` This will start a version of Grist that is like `grist-core` but with S3 snapshots enabled. To release this code to `grist-core`, it would just need to move from `ext/app` to `app` within core. I tried a lot of ways of organizing self-managed Grist, and this was what made me happiest. There are a lot of trade-offs, but here is what I was looking for: * Only OSS-code in grist-core. Adding mixed-license material there feels unfair to people already working with the repo. That said, a possible future is to move away from our private monorepo to a public mixed-licence repo, which could have the same relationship with grist-core as the monorepo has. * Minimal differences between self-managed builds and one of our existing builds, ideally hewing as close to grist-core as possible for ease of documentation, debugging, and maintenance. * Ideally, docker builds without copying files around (the new `--build-context` functionality made that possible). * Compatibility with monorepo build. Expressing dependencies of the extra code in `ext` proved tricky to do in a clean way. Yarn/npm fought me every step of the way - everything related to optional dependencies was unsatisfactory in some respect. Yarn2 is flexible but smells like it might be overreach. In the end, organizing to install non-core dependencies one directory up from the main build was a good simple trick that saved my bacon. This diff gets us to the point of building `grist-ee` images conveniently, but there isn't a public repo people can go look at to see its source. This could be generated by taking `grist-core`, adding the `ext` directory to it, and pushing to a distinct repository. I'm not in a hurry to do that, since a PR to that repo would be hard to sync with our monorepo and `grist-core`. Also, we don't have any licensing text ready for the `ext` directory. So leaving that for future work. Test Plan: manual Reviewers: georgegevoian, alexmojaki Reviewed By: georgegevoian, alexmojaki Differential Revision: https://phab.getgrist.com/D3415
2022-05-12 15:24:48 +00:00
WORKDIR /grist
COPY package.json yarn.lock /grist/
RUN yarn install --frozen-lockfile --verbose
# Install any extra node dependencies (at root level, to avoid having to wrestle
# with merging them).
COPY --from=ext / /grist/ext
RUN \
mkdir /node_modules && \
cd /grist/ext && \
{ if [ -e package.json ] ; then yarn install --frozen-lockfile --modules-folder=/node_modules --verbose ; fi }
# Build node code.
(core) add machinery for self-managed flavor of Grist Summary: Currently, we have two ways that we deliver Grist. One is grist-core, which has simple defaults and is relatively easy for third parties to deploy. The second is our internal build for our SaaS, which is the opposite. For self-managed Grist, a planned paid on-premise version of Grist, I adopt the following approach: * Use the `grist-core` build mechanism, extending it to accept an overlay of extra code if present. * Extra code is supplied in a self-contained `ext` directory, with an `ext/app` directory that is of same structure as core `app` and `stubs/app`. * The `ext` directory also contains information about extra node dependencies needed beyond that of `grist-core`. * The `ext` directory is contained within our monorepo rather than `grist-core` since it may contain material not under the Apache license. Docker builds are achieved in our monorepo by using the `--build-context` functionality to add in `ext` during the regular `grist-core` build: ``` docker buildx build --load -t gristlabs/grist-ee --build-context=ext=../ext . ``` Incremental builds in our monorepo are achieved with the `build_core.sh` helper, like: ``` buildtools/build_core.sh /tmp/self-managed cd /tmp/self-managed yarn start ``` The initial `ext` directory contains material for snapshotting to S3. If you build the docker image as above, and have S3 access, you can do something like: ``` docker run -p 8484:8484 --env GRIST_SESSION_SECRET=a-secret \ --env GRIST_DOCS_S3_BUCKET=grist-docs-test \ --env GRIST_DOCS_S3_PREFIX=self-managed \ -v $HOME/.aws:/root/.aws -it gristlabs/grist-ee ``` This will start a version of Grist that is like `grist-core` but with S3 snapshots enabled. To release this code to `grist-core`, it would just need to move from `ext/app` to `app` within core. I tried a lot of ways of organizing self-managed Grist, and this was what made me happiest. There are a lot of trade-offs, but here is what I was looking for: * Only OSS-code in grist-core. Adding mixed-license material there feels unfair to people already working with the repo. That said, a possible future is to move away from our private monorepo to a public mixed-licence repo, which could have the same relationship with grist-core as the monorepo has. * Minimal differences between self-managed builds and one of our existing builds, ideally hewing as close to grist-core as possible for ease of documentation, debugging, and maintenance. * Ideally, docker builds without copying files around (the new `--build-context` functionality made that possible). * Compatibility with monorepo build. Expressing dependencies of the extra code in `ext` proved tricky to do in a clean way. Yarn/npm fought me every step of the way - everything related to optional dependencies was unsatisfactory in some respect. Yarn2 is flexible but smells like it might be overreach. In the end, organizing to install non-core dependencies one directory up from the main build was a good simple trick that saved my bacon. This diff gets us to the point of building `grist-ee` images conveniently, but there isn't a public repo people can go look at to see its source. This could be generated by taking `grist-core`, adding the `ext` directory to it, and pushing to a distinct repository. I'm not in a hurry to do that, since a PR to that repo would be hard to sync with our monorepo and `grist-core`. Also, we don't have any licensing text ready for the `ext` directory. So leaving that for future work. Test Plan: manual Reviewers: georgegevoian, alexmojaki Reviewed By: georgegevoian, alexmojaki Differential Revision: https://phab.getgrist.com/D3415
2022-05-12 15:24:48 +00:00
COPY tsconfig.json /grist
COPY tsconfig-ext.json /grist
COPY test/tsconfig.json /grist/test/tsconfig.json
COPY app /grist/app
COPY stubs /grist/stubs
COPY buildtools /grist/buildtools
RUN yarn run build:prod
################################################################################
## Python collection stage
################################################################################
# Fetch python3.9 and python2.7
FROM python:3.9-slim-buster as collector
# Install all python dependencies.
ADD sandbox/requirements.txt requirements.txt
ADD sandbox/requirements3.txt requirements3.txt
RUN \
apt update && \
2022-04-03 23:43:29 +00:00
apt install -y --no-install-recommends python2 python-pip python-setuptools \
build-essential libxml2-dev libxslt-dev python-dev zlib1g-dev && \
pip2 install wheel && \
pip2 install -r requirements.txt && \
pip3 install -r requirements3.txt
################################################################################
## Sandbox collection stage
################################################################################
# Fetch gvisor-based sandbox. Note, to enable it to run within default
# unprivileged docker, layers of protection that require privilege have
# been stripped away, see https://github.com/google/gvisor/issues/4371
FROM gristlabs/gvisor-unprivileged:buster as sandbox
################################################################################
## Run-time stage
################################################################################
# Now, start preparing final image.
FROM node:14-buster-slim
# Install libexpat1, libsqlite3-0 for python3 library binary dependencies.
# Install pgrep for managing gvisor processes.
RUN \
apt-get update && \
apt-get install -y --no-install-recommends libexpat1 libsqlite3-0 procps && \
rm -rf /var/lib/apt/lists/*
# Keep all storage user may want to persist in a distinct directory
RUN mkdir -p /persist/docs
# Copy node files.
(core) add machinery for self-managed flavor of Grist Summary: Currently, we have two ways that we deliver Grist. One is grist-core, which has simple defaults and is relatively easy for third parties to deploy. The second is our internal build for our SaaS, which is the opposite. For self-managed Grist, a planned paid on-premise version of Grist, I adopt the following approach: * Use the `grist-core` build mechanism, extending it to accept an overlay of extra code if present. * Extra code is supplied in a self-contained `ext` directory, with an `ext/app` directory that is of same structure as core `app` and `stubs/app`. * The `ext` directory also contains information about extra node dependencies needed beyond that of `grist-core`. * The `ext` directory is contained within our monorepo rather than `grist-core` since it may contain material not under the Apache license. Docker builds are achieved in our monorepo by using the `--build-context` functionality to add in `ext` during the regular `grist-core` build: ``` docker buildx build --load -t gristlabs/grist-ee --build-context=ext=../ext . ``` Incremental builds in our monorepo are achieved with the `build_core.sh` helper, like: ``` buildtools/build_core.sh /tmp/self-managed cd /tmp/self-managed yarn start ``` The initial `ext` directory contains material for snapshotting to S3. If you build the docker image as above, and have S3 access, you can do something like: ``` docker run -p 8484:8484 --env GRIST_SESSION_SECRET=a-secret \ --env GRIST_DOCS_S3_BUCKET=grist-docs-test \ --env GRIST_DOCS_S3_PREFIX=self-managed \ -v $HOME/.aws:/root/.aws -it gristlabs/grist-ee ``` This will start a version of Grist that is like `grist-core` but with S3 snapshots enabled. To release this code to `grist-core`, it would just need to move from `ext/app` to `app` within core. I tried a lot of ways of organizing self-managed Grist, and this was what made me happiest. There are a lot of trade-offs, but here is what I was looking for: * Only OSS-code in grist-core. Adding mixed-license material there feels unfair to people already working with the repo. That said, a possible future is to move away from our private monorepo to a public mixed-licence repo, which could have the same relationship with grist-core as the monorepo has. * Minimal differences between self-managed builds and one of our existing builds, ideally hewing as close to grist-core as possible for ease of documentation, debugging, and maintenance. * Ideally, docker builds without copying files around (the new `--build-context` functionality made that possible). * Compatibility with monorepo build. Expressing dependencies of the extra code in `ext` proved tricky to do in a clean way. Yarn/npm fought me every step of the way - everything related to optional dependencies was unsatisfactory in some respect. Yarn2 is flexible but smells like it might be overreach. In the end, organizing to install non-core dependencies one directory up from the main build was a good simple trick that saved my bacon. This diff gets us to the point of building `grist-ee` images conveniently, but there isn't a public repo people can go look at to see its source. This could be generated by taking `grist-core`, adding the `ext` directory to it, and pushing to a distinct repository. I'm not in a hurry to do that, since a PR to that repo would be hard to sync with our monorepo and `grist-core`. Also, we don't have any licensing text ready for the `ext` directory. So leaving that for future work. Test Plan: manual Reviewers: georgegevoian, alexmojaki Reviewed By: georgegevoian, alexmojaki Differential Revision: https://phab.getgrist.com/D3415
2022-05-12 15:24:48 +00:00
COPY --from=builder /node_modules /node_modules
COPY --from=builder /grist/node_modules /grist/node_modules
COPY --from=builder /grist/_build /grist/_build
COPY --from=builder /grist/static /grist/static-built
# Copy python files.
COPY --from=collector /usr/bin/python2.7 /usr/bin/python2.7
COPY --from=collector /usr/lib/python2.7 /usr/lib/python2.7
COPY --from=collector /usr/local/lib/python2.7 /usr/local/lib/python2.7
COPY --from=collector /usr/local/bin/python3.9 /usr/bin/python3.9
COPY --from=collector /usr/local/lib/python3.9 /usr/local/lib/python3.9
COPY --from=collector /usr/local/lib/libpython3.9.* /usr/local/lib/
# Set default to python3
RUN \
ln -s /usr/bin/python3.9 /usr/bin/python && \
ln -s /usr/bin/python3.9 /usr/bin/python3 && \
ldconfig
# Copy runsc.
COPY --from=sandbox /runsc /usr/bin/runsc
# Add files needed for running server.
(core) add machinery for self-managed flavor of Grist Summary: Currently, we have two ways that we deliver Grist. One is grist-core, which has simple defaults and is relatively easy for third parties to deploy. The second is our internal build for our SaaS, which is the opposite. For self-managed Grist, a planned paid on-premise version of Grist, I adopt the following approach: * Use the `grist-core` build mechanism, extending it to accept an overlay of extra code if present. * Extra code is supplied in a self-contained `ext` directory, with an `ext/app` directory that is of same structure as core `app` and `stubs/app`. * The `ext` directory also contains information about extra node dependencies needed beyond that of `grist-core`. * The `ext` directory is contained within our monorepo rather than `grist-core` since it may contain material not under the Apache license. Docker builds are achieved in our monorepo by using the `--build-context` functionality to add in `ext` during the regular `grist-core` build: ``` docker buildx build --load -t gristlabs/grist-ee --build-context=ext=../ext . ``` Incremental builds in our monorepo are achieved with the `build_core.sh` helper, like: ``` buildtools/build_core.sh /tmp/self-managed cd /tmp/self-managed yarn start ``` The initial `ext` directory contains material for snapshotting to S3. If you build the docker image as above, and have S3 access, you can do something like: ``` docker run -p 8484:8484 --env GRIST_SESSION_SECRET=a-secret \ --env GRIST_DOCS_S3_BUCKET=grist-docs-test \ --env GRIST_DOCS_S3_PREFIX=self-managed \ -v $HOME/.aws:/root/.aws -it gristlabs/grist-ee ``` This will start a version of Grist that is like `grist-core` but with S3 snapshots enabled. To release this code to `grist-core`, it would just need to move from `ext/app` to `app` within core. I tried a lot of ways of organizing self-managed Grist, and this was what made me happiest. There are a lot of trade-offs, but here is what I was looking for: * Only OSS-code in grist-core. Adding mixed-license material there feels unfair to people already working with the repo. That said, a possible future is to move away from our private monorepo to a public mixed-licence repo, which could have the same relationship with grist-core as the monorepo has. * Minimal differences between self-managed builds and one of our existing builds, ideally hewing as close to grist-core as possible for ease of documentation, debugging, and maintenance. * Ideally, docker builds without copying files around (the new `--build-context` functionality made that possible). * Compatibility with monorepo build. Expressing dependencies of the extra code in `ext` proved tricky to do in a clean way. Yarn/npm fought me every step of the way - everything related to optional dependencies was unsatisfactory in some respect. Yarn2 is flexible but smells like it might be overreach. In the end, organizing to install non-core dependencies one directory up from the main build was a good simple trick that saved my bacon. This diff gets us to the point of building `grist-ee` images conveniently, but there isn't a public repo people can go look at to see its source. This could be generated by taking `grist-core`, adding the `ext` directory to it, and pushing to a distinct repository. I'm not in a hurry to do that, since a PR to that repo would be hard to sync with our monorepo and `grist-core`. Also, we don't have any licensing text ready for the `ext` directory. So leaving that for future work. Test Plan: manual Reviewers: georgegevoian, alexmojaki Reviewed By: georgegevoian, alexmojaki Differential Revision: https://phab.getgrist.com/D3415
2022-05-12 15:24:48 +00:00
ADD package.json /grist/package.json
ADD ormconfig.js /grist/ormconfig.js
ADD bower_components /grist/bower_components
ADD sandbox /grist/sandbox
ADD plugins /grist/plugins
ADD static /grist/static
# Finalize static directory
RUN \
mv /grist/static-built/* /grist/static && \
rmdir /grist/static-built
WORKDIR /grist
# Set some default environment variables to give a setup that works out of the box when
# started as:
# docker run -p 8484:8484 -it <image>
# Variables will need to be overridden for other setups.
#
# GRIST_SANDBOX_FLAVOR is set to unsandboxed by default, because it
# appears that the services people use to run docker containers have
# a wide variety of security settings and the functionality needed for
# sandboxing may not be possible in every case. For default docker
# settings, you can get sandboxing as follows:
# docker run --env GRIST_SANDBOX_FLAVOR=gvisor -p 8484:8484 -it <image>
#
ENV \
PYTHON_VERSION_ON_CREATION=3 \
GRIST_ORG_IN_PATH=true \
GRIST_HOST=0.0.0.0 \
GRIST_SINGLE_PORT=true \
GRIST_SERVE_SAME_ORIGIN=true \
GRIST_DATA_DIR=/persist/docs \
GRIST_INST_DIR=/persist \
GRIST_SESSION_COOKIE=grist_core \
GVISOR_FLAGS="-unprivileged -ignore-cgroups" \
GRIST_SANDBOX_FLAVOR=unsandboxed \
TYPEORM_DATABASE=/persist/home.sqlite3
EXPOSE 8484
CMD ./sandbox/run.sh