This page describes how gsutil uses a boto configuration file and provides a collaboration example that uses the file. The boto configuration file is also used by boto, which is the Amazon S3 SDK for Python.
Configuration file overview
The boto configuration file contains values that control how gsutil behaves. For
prefer_api variable determines which API gsutil preferentially
uses. Boto configuration file variables can be changed by editing the
configuration file directly. While most users won't need to edit these
variables, those do typically do so for one of the following reasons:
- Setting up gsutil to work through a proxy.
- Using customer-managed or customer-supplied encryption keys.
- Performing specialized customization of global gsutil behavior.
The default location for the boto configuration file is in the user home
directory, ~/.boto, for Linux and macOS, and in %HOMEDRIVE%%HOMEPATH%,
for Windows. You can get the location of the configuration file by running
gsutil version -l.
You can override where gsutil expects to find the config file by setting the
BOTO_CONFIG environment variable. You can also set up a path of boto
configuration files to load by setting the
BOTO_PATH environment variable
: delimited path (or
; for Windows). For example, setting the
BOTO_PATH environment variable to:
causes gsutil to load each configuration file found in the path in order. This is useful if you want to set up some shared configuration state among many users. See the example of using the configuration file for such a data sharing and collaboration scenario.
The configuration file contains a number of sections:
[OAuth2]. The following are the currently defined
configuration settings, broken down by section. Their use is documented in the
boto configuration file itself, in comments preceding each setting:
[Credentials] aws_access_key_id aws_secret_access_key gs_access_key_id gs_host gs_host_header gs_json_host gs_json_host_header gs_json_port gs_oauth2_refresh_token gs_port gs_secret_access_key gs_service_client_id gs_service_key_file gs_service_key_file_password s3_host s3_host_header s3_port [Boto] proxy proxy_type proxy_port proxy_user proxy_pass proxy_rdns http_socket_timeout ca_certificates_file https_validate_certificates debug max_retry_delay num_retries [GoogleCompute] service_account [GSUtil] check_hashes content_language decryption_key1 ... 100 default_api_version default_project_id disable_analytics_prompt encryption_key json_api_version max_upload_compression_buffer_size parallel_composite_upload_component_size parallel_composite_upload_threshold sliced_object_download_component_size sliced_object_download_max_components sliced_object_download_threshold parallel_process_count parallel_thread_count gzip_compression_level prefer_api resumable_threshold resumable_tracker_dir (deprecated in 4.6, use state_dir) rsync_buffer_lines software_update_check_period state_dir tab_completion_time_logs tab_completion_timeout task_estimation_threshold test_cmd_regional_bucket_location test_notification_url use_magicfile test_hmac_service_account test_hmac_alt_service_account test_hmac_list_service_account [OAuth2] client_id client_secret oauth2_refresh_retries provider_authorization_uri provider_label provider_token_uri token_cache
When you edit the file, be careful not to mis-edit any of the setting names,
gs_access_key_id, and do not remove the section delimiters, such as
Updating to the latest configuration file
New configuration controllable features are added to the boto configuration file
over time, but most gsutil users create a configuration file once and then
keep it for a long time. This means that new features aren't apparent when you
update to a newer version of gsutil. If you want to get the latest configuration
file, which includes the latest settings and documentation, rename your current
file (e.g., to
gsutil config or
gcloud init, and then
transfer any configuration settings you want to keep from your old file into the
newly created file. Note, however, that if you're using OAuth2 credentials and
you go back through the OAuth2 configuration process, doing so invalidates your
previous OAuth2 credentials.
Example using the configuration file
In this example, a small company wants to use Cloud Storage as a storage system for their employees. As the IT administrator, you create a project in the Google Cloud Console and create buckets for each employee. To make it easier for employees to use Cloud Storage, you want to create and store company-wide settings, such as a proxy configuration and parallel composite upload thresholds, in a central file that employees can point to in their BOTO configuration path. This eliminates the need for each employee to set the shared parts of the configuration manually and allows you, as the administrator, to easily change these shared configurations if necessary.
To accomplish this, perform the following steps:
Create a central boto configuration file that is readable by all employees.
This can be done by using
gcloud initwhen gsutil is installed as part of the Google Cloud SDK.
The boto configuration file might contain, for example:
[Boto] proxy = yourproxy.com proxy_port = 8080 proxy_type = http [GSUtil] parallel_composite_upload_threshold = 150M
Instruct employees to install the Google Cloud SDK.
During installation, employees need to specify the project ID that the company is using. They also need to generate individual authentication credentials because you should never share authentication credentials centrally.
Instruct employees to add a BOTO_PATH environment variable.
The BOTO_PATH environment variable lists the path of the centrally located configuration file, followed by the employee's local configuration file. For example, if the central configuration file is located in the directory
centralhub/, then for user
jane, the BOTO_PATH environment variable might be:
When employees run gsutil, they will automatically use the configuration specified in the central boto file. If necessary, the administrator can change the proxy settings, parallel composite upload threshold, and other settings in the central configuration file and have the changes reflected for all employees using the central configuration file.