diff --git a/docs/self-hosted/external-services/object-storage.mdx b/docs/self-hosted/external-services/object-storage.mdx index d78b4591a..c8aff7196 100644 --- a/docs/self-hosted/external-services/object-storage.mdx +++ b/docs/self-hosted/external-services/object-storage.mdx @@ -1,118 +1,230 @@ # Using a managed object storage service (S3 or GCS) -By default, Sourcegraph will use a `sourcegraph/blobstore` server bundled with the instance to temporarily store [code graph indexes](../../code-navigation/precise-code-navigation) uploaded by users as well as the results of [search jobs](../../code-search/types/search-jobs). +Object storage is used by various Sourcegraph features, for example to store [code graph indexes](../../code-navigation/precise-code-navigation) uploaded by users or the results of [search jobs](../../code-search/types/search-jobs). -You can alternatively configure your instance to instead store this data in an S3 or GCS bucket. Doing so may decrease your hosting costs as persistent volumes are often more expensive than the same storage space in an object store service. +By default, Sourcegraph will use a `sourcegraph/blobstore` server bundled with the instance. This is intended as a temporary measure: we recommend that administrators configure self-hosted Sourcegraph to store this data in an [AWS S3](https://aws.amazon.com/s3/) or [Google Cloud Storage (GCS)](https://cloud.google.com/storage) bucket following the guidance below. Doing so may decrease your hosting costs as persistent volumes are often more expensive than the same storage space in an object store service. + + + Starting in [Sourcegraph + 7.2](https://sourcegraph.com/changelog/releases/7.2), new instances only + need to configure the [Sourcegraph bucket](#sourcegraph-bucket), and + Sourcegraph will use that single bucket for all features. Instances + provisioned before 7.2 can continue to use their existing buckets, but the + new bucket is still required. + + +## Sourcegraph bucket + + + Self-hosted Sourcegraph instances using S3 or GCS object storage provisioned + before [Sourcegraph 7.2](https://sourcegraph.com/changelog/releases/7.2) + should provision an additional bucket following the guidance below. + Sourcegraph will report a warning when this bucket is not present, as it + will become required for new features in a future release. No action is + required if you are using the default `sourcegraph/blobstore`. + + +The Sourcegraph bucket is intended to be the single bucket for new Sourcegraph features. Instead of creating one bucket per feature, new features store objects under namespaced key prefixes within this bucket. + +Existing bucket configuration for [code graph indexes](#code-graph-indexes) and [search jobs](#search-job-results) remain in use. This change ensures future features can be enabled without requiring a new bucket for each feature. + +New instances deployed using [Sourcegraph 7.2](https://sourcegraph.com/changelog/releases/7.2) or later can choose to _only_ provision the Sourcegraph bucket: this bucket can be used for both [code graph indexes](#code-graph-indexes) and [search jobs](#search-job-results), if not explicit configuration is provided for those features. + +### Using GCS for the Sourcegraph bucket + +Set the following environment variables to target a GCS bucket for shared Sourcegraph uploads. + +- `SOURCEGRAPH_UPLOAD_BACKEND=GCS` +- `SOURCEGRAPH_UPLOAD_BUCKET=` +- `SOURCEGRAPH_UPLOAD_GCP_PROJECT_ID=` +- `SOURCEGRAPH_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE=` (optional) +- `SOURCEGRAPH_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE_CONTENT=<{"my": "content"}>` (optional) + +If you are running on GKE with Workload Identity, or otherwise relying on +Application Default Credentials, you can omit the GCS credentials file +variables. Grant the `roles/storage.objectAdmin` role to the service accounts used by +the `frontend`, `worker`, `precise-code-intel-worker`, `gitserver`, and +`searcher` services in a GKE environment. + +### Using S3 for the Sourcegraph bucket + +Set the following environment variables to target an S3 bucket for shared Sourcegraph uploads. + +- `SOURCEGRAPH_UPLOAD_BACKEND=S3` +- `SOURCEGRAPH_UPLOAD_BUCKET=` +- `SOURCEGRAPH_UPLOAD_AWS_REGION=us-east-1` +- `SOURCEGRAPH_UPLOAD_AWS_ENDPOINT=https://s3.us-east-1.amazonaws.com` +- `SOURCEGRAPH_UPLOAD_AWS_ACCESS_KEY_ID=` +- `SOURCEGRAPH_UPLOAD_AWS_SECRET_ACCESS_KEY=` +- `SOURCEGRAPH_UPLOAD_AWS_SESSION_TOKEN=` (optional) +- `SOURCEGRAPH_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` (optional; set to use EC2 metadata API over static credentials) +- `SOURCEGRAPH_UPLOAD_AWS_USE_PATH_STYLE=false` (optional) + + + If a non-default region is supplied, ensure that the subdomain of the + endpoint URL (_the `AWS_ENDPOINT` value_) matches the target region. + + +You don't need to set the `SOURCEGRAPH_UPLOAD_AWS_ACCESS_KEY_ID` environment +variable when using `SOURCEGRAPH_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` +because role credentials will be automatically resolved. Attach the IAM role +to the EC2 instances hosting the `frontend`, `worker`, +`precise-code-intel-worker`, `gitserver`, and `searcher` containers in a +multi-node environment. + +### Automatically provision the Sourcegraph bucket + +Most deployments should provision this bucket directly in their cloud provider and leave this disabled. If you would like to allow your Sourcegraph instance to manage the target bucket configuration, set the following environment variable: + + + This requires additional bucket-management permissions from your configured + storage vendor (AWS or GCP). + + +- `SOURCEGRAPH_UPLOAD_MANAGE_BUCKET=true` ## Code Graph Indexes To target a managed object storage service for storing [code graph index uploads](../../code-navigation/precise-code-navigation), you will need to set a handful of environment variables for configuration and authentication to the target service. -- If you are running a `sourcegraph/server` deployment, set the environment variables on the server container -- If you are running via Docker-compose or Kubernetes, set the environment variables on the `frontend`, `worker`, and `precise-code-intel-worker` containers + + Starting in [Sourcegraph + 7.2](https://sourcegraph.com/changelog/releases/7.2), new instances only + need to configure the [Sourcegraph bucket](#sourcegraph-bucket), and + Sourcegraph will use that single bucket for all features. If a separate + bucket is needed for Code Graph Indexes, that can still be configured, but + we recommend using one bucket. + -### Using S3 +- If you are running a `sourcegraph/server` deployment, set the environment variables on the server container +- If you are running via Docker-compose or Kubernetes, set the environment variables on the `frontend`, `worker`, and `precise-code-intel-worker` containers + +### Using S3 for the Code Graph Indexes bucket To target an S3 bucket you've already provisioned, set the following environment variables. Authentication can be done through [an access and secret key pair](https://docs.aws.amazon.com/general/latest/gr/aws-sec-cred-types.html#access-keys-and-secret-access-keys) (and optional session token), or via the EC2 metadata API. - {' '} - Never commit AWS access keys in Git. You should consider using a secret handling - service offered by your cloud provider.{' '} + Never commit AWS access keys in Git. You should consider using a secret + handling service offered by your cloud provider. -- `PRECISE_CODE_INTEL_UPLOAD_BACKEND=S3` -- `PRECISE_CODE_INTEL_UPLOAD_BUCKET=` -- `PRECISE_CODE_INTEL_UPLOAD_AWS_ENDPOINT=https://s3.us-east-1.amazonaws.com` -- `PRECISE_CODE_INTEL_UPLOAD_AWS_ACCESS_KEY_ID=` -- `PRECISE_CODE_INTEL_UPLOAD_AWS_SECRET_ACCESS_KEY=` -- `PRECISE_CODE_INTEL_UPLOAD_AWS_SESSION_TOKEN=` (optional) -- `PRECISE_CODE_INTEL_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` (optional; set to use EC2 metadata API over static credentials) -- `PRECISE_CODE_INTEL_UPLOAD_AWS_REGION=us-east-1` (default) +- `PRECISE_CODE_INTEL_UPLOAD_BACKEND=S3` +- `PRECISE_CODE_INTEL_UPLOAD_BUCKET=` +- `PRECISE_CODE_INTEL_UPLOAD_AWS_REGION=us-east-1` +- `PRECISE_CODE_INTEL_UPLOAD_AWS_ENDPOINT=https://s3.us-east-1.amazonaws.com` +- `PRECISE_CODE_INTEL_UPLOAD_AWS_ACCESS_KEY_ID=` +- `PRECISE_CODE_INTEL_UPLOAD_AWS_SECRET_ACCESS_KEY=` +- `PRECISE_CODE_INTEL_UPLOAD_AWS_SESSION_TOKEN=` (optional) +- `PRECISE_CODE_INTEL_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` (optional; set to use EC2 metadata API over static credentials) - {' '} - If a non-default region is supplied, ensure that the subdomain of the endpoint - URL (_the `AWS_ENDPOINT` value_) matches the target region.{' '} + If a non-default region is supplied, ensure that the subdomain of the + endpoint URL (_the `AWS_ENDPOINT` value_) matches the target region. - - {' '} - You don't need to set the `PRECISE_CODE_INTEL_UPLOAD_AWS_ACCESS_KEY_ID` environment - variable when using `PRECISE_CODE_INTEL_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` - because role credentials will be automatically resolved. Attach the IAM role - to the EC2 instances hosting the `frontend`, `worker`, and `precise-code-intel-worker` - containers in a multi-node environment.{' '} - +You don't need to set the `PRECISE_CODE_INTEL_UPLOAD_AWS_ACCESS_KEY_ID` +environment variable when using +`PRECISE_CODE_INTEL_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` because role +credentials will be automatically resolved. Attach the IAM role to the EC2 +instances hosting the `frontend`, `worker`, and `precise-code-intel-worker` +containers in a multi-node environment. + +### Using GCS for the Code Graph Indexes bucket -### Using GCS +To target a GCS bucket you've already provisioned, set the following environment variables. -To target a GCS bucket you've already provisioned, set the following environment variables. Authentication is done through a service account key, supplied as either a path to a volume-mounted file, or the contents read in as an environment variable payload. +- `PRECISE_CODE_INTEL_UPLOAD_BACKEND=GCS` +- `PRECISE_CODE_INTEL_UPLOAD_BUCKET=` +- `PRECISE_CODE_INTEL_UPLOAD_GCP_PROJECT_ID=` +- `PRECISE_CODE_INTEL_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE=` (optional) +- `PRECISE_CODE_INTEL_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE_CONTENT=<{"my": "content"}>` (optional) -- `PRECISE_CODE_INTEL_UPLOAD_BACKEND=GCS` -- `PRECISE_CODE_INTEL_UPLOAD_BUCKET=` -- `PRECISE_CODE_INTEL_UPLOAD_GCP_PROJECT_ID=` -- `PRECISE_CODE_INTEL_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE=` -- `PRECISE_CODE_INTEL_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE_CONTENT=<{"my": "content"}>` +If you are running on GKE with Workload Identity, or otherwise relying on +Application Default Credentials, you can omit the GCS credentials file +variables. Grant the `roles/storage.objectAdmin` role to the service accounts used by +the `frontend`, `worker`, and `precise-code-intel-worker` services in a GKE environment. -### Provisioning buckets +### Automatically provision the Code Graph Indexes bucket If you would like to allow your Sourcegraph instance to control the creation and lifecycle configuration management of the target buckets, set the following environment variables: -- `PRECISE_CODE_INTEL_UPLOAD_MANAGE_BUCKET=true` -- `PRECISE_CODE_INTEL_UPLOAD_TTL=168h` (default) + + This requires additional bucket-management permissions from your configured + storage vendor (AWS or GCP). + + +- `PRECISE_CODE_INTEL_UPLOAD_MANAGE_BUCKET=true` +- `PRECISE_CODE_INTEL_UPLOAD_TTL=168h` (default) -## Search Job Results +## Search Job results -To target a third party managed object storage service for storing [search job results](../../code-search/types/search-jobs), you must set a handful of environment variables for configuration and authentication to the target service. +To target a third party managed object storage service for storing [Search Job results](../../code-search/types/search-jobs), you must set a handful of environment variables for configuration and authentication to the target service. + + + Starting in [Sourcegraph + 7.2](https://sourcegraph.com/changelog/releases/7.2), new instances only + need to configure the [Sourcegraph bucket](#sourcegraph-bucket), and + Sourcegraph will use that single bucket for all features. If a separate + bucket is needed for Search Job results, that can still be configured, but + we recommend using one bucket. + -- If you are running a `sourcegraph/server` deployment, set the environment variables on the server container -- If you are running via Docker-compose or Kubernetes, set the environment variables on the `frontend` and `worker` containers +- If you are running a `sourcegraph/server` deployment, set the environment variables on the server container +- If you are running via Docker-compose or Kubernetes, set the environment variables on the `frontend` and `worker` containers -### Using S3 +### Using S3 for the Search Job results bucket Set the following environment variables to target an S3 bucket you've already provisioned. Authentication can be done through [an access and secret key pair](https://docs.aws.amazon.com/general/latest/gr/aws-sec-cred-types.html#access-keys-and-secret-access-keys) (and optionally through session token) or via the EC2 metadata API. - {' '} - Never commit AWS access keys in Git. You should consider using a secret handling - service offered by your cloud provider. + Never commit AWS access keys in Git. You should consider using a secret + handling service offered by your cloud provider. -- `SEARCH_JOBS_UPLOAD_BACKEND=S3` -- `SEARCH_JOBS_UPLOAD_BUCKET=` -- `SEARCH_JOBS_UPLOAD_AWS_ENDPOINT=https://s3.us-east-1.amazonaws.com` -- `SEARCH_JOBS_UPLOAD_AWS_ACCESS_KEY_ID=` -- `SEARCH_JOBS_UPLOAD_AWS_SECRET_ACCESS_KEY=` -- `SEARCH_JOBS_UPLOAD_AWS_SESSION_TOKEN=` (optional) -- `SEARCH_JOBS_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` (optional; set to use EC2 metadata API over static credentials) -- `SEARCH_JOBS_UPLOAD_AWS_REGION=us-east-1` (default) +- `SEARCH_JOBS_UPLOAD_BACKEND=S3` +- `SEARCH_JOBS_UPLOAD_BUCKET=` +- `SEARCH_JOBS_UPLOAD_AWS_REGION=us-east-1` +- `SEARCH_JOBS_UPLOAD_AWS_ENDPOINT=https://s3.us-east-1.amazonaws.com` +- `SEARCH_JOBS_UPLOAD_AWS_ACCESS_KEY_ID=` +- `SEARCH_JOBS_UPLOAD_AWS_SECRET_ACCESS_KEY=` +- `SEARCH_JOBS_UPLOAD_AWS_SESSION_TOKEN=` (optional) +- `SEARCH_JOBS_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` (optional; set to use EC2 metadata API over static credentials) - {' '} - If a non-default region is supplied, ensure that the subdomain of the endpoint - URL (the `AWS_ENDPOINT` value) matches the target region. + If a non-default region is supplied, ensure that the subdomain of the + endpoint URL (_the `AWS_ENDPOINT` value_) matches the target region. - {' '} You don't need to set the `SEARCH_JOBS_UPLOAD_AWS_ACCESS_KEY_ID` environment - variable when using `SEARCH_JOBS_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` because - role credentials will be automatically resolved. + variable when using `SEARCH_JOBS_UPLOAD_AWS_USE_EC2_ROLE_CREDENTIALS=true` + because role credentials will be automatically resolved. Attach the IAM role + to the EC2 instances hosting the `frontend` and `worker` containers in a + multi-node environment. -### Using GCS +### Using GCS for the Search Job results bucket + +Set the following environment variables to target a GCS bucket you've already provisioned. -Set the following environment variables to target a GCS bucket you've already provisioned. Authentication is done through a service account key, either as a path to a volume-mounted file or the contents read in as an environment variable payload. +- `SEARCH_JOBS_UPLOAD_BACKEND=GCS` +- `SEARCH_JOBS_UPLOAD_BUCKET=` +- `SEARCH_JOBS_UPLOAD_GCP_PROJECT_ID=` +- `SEARCH_JOBS_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE=` (optional) +- `SEARCH_JOBS_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE_CONTENT=<{"my": "content"}>` (optional) -- `SEARCH_JOBS_UPLOAD_BACKEND=GCS` -- `SEARCH_JOBS_UPLOAD_BUCKET=` -- `SEARCH_JOBS_UPLOAD_GCP_PROJECT_ID=` -- `SEARCH_JOBS_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE=` -- `SEARCH_JOBS_UPLOAD_GOOGLE_APPLICATION_CREDENTIALS_FILE_CONTENT=<{"my": "content"}>` +If you are running on GKE with Workload Identity, or otherwise relying on +Application Default Credentials, you can omit the GCS credentials file +variables. Grant the `roles/storage.objectAdmin` role to the service accounts used by +the `frontend` and `worker` services in a GKE environment. -### Provisioning buckets +### Automatically provision the Search Job results bucket If you would like to allow your Sourcegraph instance to control the creation and lifecycle configuration management of the target buckets, set the following environment variables: -- `SEARCH_JOBS_UPLOAD_MANAGE_BUCKET=true` + + This requires additional bucket-management permissions from your configured + storage vendor (AWS or GCP). + + +- `SEARCH_JOBS_UPLOAD_MANAGE_BUCKET=true`