1 # Copyright (C) The Arvados Authors. All rights reserved.
3 # SPDX-License-Identifier: AGPL-3.0
5 # Do not use this file for site configuration. Create application.yml
6 # instead (see application.yml.example).
8 # The order of precedence is:
9 # 1. config/environments/{RAILS_ENV}.rb (deprecated)
10 # 2. Section in application.yml corresponding to RAILS_ENV (e.g., development)
11 # 3. Section in application.yml called "common"
12 # 4. Section in application.default.yml corresponding to RAILS_ENV
13 # 5. Section in application.default.yml called "common"
17 ### Essential site configuration
20 # The prefix used for all database identifiers to identify the record as
21 # originating from this site. Must be exactly 5 alphanumeric characters
22 # (lowercase ASCII letters and digits).
25 # secret_token is a string of alphanumeric characters used by Rails
26 # to sign session tokens. IMPORTANT: This is a site secret. It
27 # should be at least 50 characters.
30 # blob_signing_key is a string of alphanumeric characters used to
31 # generate permission signatures for Keep locators. It must be
32 # identical to the permission key given to Keep. IMPORTANT: This is
33 # a site secret. It should be at least 50 characters.
35 # Modifying blob_signing_key will invalidate all existing
36 # signatures, which can cause programs to fail (e.g., arv-put,
37 # arv-get, and Crunch jobs). To avoid errors, rotate keys only when
38 # no such processes are running.
41 # These settings are provided by your OAuth2 provider (e.g.,
47 # If this is not false, HTML requests at the API server's root URL
48 # are redirected to this location, and it is provided in the text of
49 # user activation notification email messages to remind them where
51 workbench_address: false
53 # Client-facing URI for websocket service. Nginx should be
54 # configured to proxy this URI to arvados-ws; see
55 # http://doc.arvados.org/install/install-ws.html
57 # If websocket_address is false (which is the default), no websocket
58 # server will be advertised to clients. This configuration is not
62 #websocket_address: wss://ws.zzzzz.arvadosapi.com/websocket
63 websocket_address: false
65 # Maximum number of websocket connections allowed
66 websocket_max_connections: 500
68 # Maximum number of events a single connection can be backlogged
69 websocket_max_notify_backlog: 1000
71 # Maximum number of subscriptions a single websocket connection can have
73 websocket_max_filters: 10
75 # Git repositories must be readable by api server, or you won't be
76 # able to submit crunch jobs. To pass the test suites, put a clone
77 # of the arvados tree in {git_repositories_dir}/arvados.git or
78 # {git_repositories_dir}/arvados/.git
79 git_repositories_dir: /var/lib/arvados/git/repositories
81 # This is a (bare) repository that stores commits used in jobs. When a job
82 # runs, the source commits are first fetched into this repository, then this
83 # repository is used to deploy to compute nodes. This should NOT be a
84 # subdirectory of {git_repositiories_dir}.
85 git_internal_dir: /var/lib/arvados/internal.git
87 # Default replication level for collections. This is used when a
88 # collection's replication_desired attribute is nil.
89 default_collection_replication: 2
93 ### Overriding default advertised hostnames/URLs
96 # If not false, this is the hostname, port, and protocol that will be used
97 # for root_url and advertised in the discovery document. By default, use
98 # the default Rails logic for deciding on a hostname.
103 # Base part of SSH git clone url given with repository resources. If
104 # true, the default "git@git.(uuid_prefix).arvadosapi.com:" is
105 # used. If false, SSH clone URLs are not advertised. Include a
106 # trailing ":" or "/" if needed: it will not be added automatically.
107 git_repo_ssh_base: true
109 # Base part of HTTPS git clone urls given with repository
110 # resources. This is expected to be an arv-git-httpd service which
111 # accepts API tokens as HTTP-auth passwords. If true, the default
112 # "https://git.(uuid_prefix).arvadosapi.com/" is used. If false,
113 # HTTPS clone URLs are not advertised. Include a trailing ":" or "/"
114 # if needed: it will not be added automatically.
115 git_repo_https_base: true
119 ### New user and & email settings
122 # Config parameters to automatically setup new users. If enabled,
123 # this users will be able to self-activate. Enable this if you want
124 # to run an open instance where anyone can create an account and use
125 # the system without requiring manual approval.
127 # The params auto_setup_new_users_with_* are meaningful only when auto_setup_new_users is turned on.
128 # auto_setup_name_blacklist is a list of usernames to be blacklisted for auto setup.
129 auto_setup_new_users: false
130 auto_setup_new_users_with_vm_uuid: false
131 auto_setup_new_users_with_repository: false
132 auto_setup_name_blacklist: [arvados, git, gitolite, gitolite-admin, root, syslog]
134 # When new_users_are_active is set to true, new users will be active
135 # immediately. This skips the "self-activate" step which enforces
136 # user agreements. Should only be enabled for development.
137 new_users_are_active: false
139 # The e-mail address of the user you would like to become marked as an admin
140 # user on their first login.
141 # In the default configuration, authentication happens through the Arvados SSO
142 # server, which uses OAuth2 against Google's servers, so in that case this
143 # should be an address associated with a Google account.
144 auto_admin_user: false
146 # If auto_admin_first_user is set to true, the first user to log in when no
147 # other admin users exist will automatically become an admin user.
148 auto_admin_first_user: false
150 # Email address to notify whenever a user creates a profile for the
152 user_profile_notification_address: false
154 admin_notifier_email_from: arvados@example.com
155 email_subject_prefix: "[ARVADOS] "
156 user_notifier_email_from: arvados@example.com
157 new_user_notification_recipients: [ ]
158 new_inactive_user_notification_recipients: [ ]
162 ### Limits, timeouts and durations
165 # Lifetime (in seconds) of blob permission signatures generated by
166 # the API server. This determines how long a client can take (after
167 # retrieving a collection record) to retrieve the collection data
168 # from Keep. If the client needs more time than that (assuming the
169 # collection still has the same content and the relevant user/token
170 # still has permission) the client can retrieve the collection again
171 # to get fresh signatures.
173 # This must be exactly equal to the -blob-signature-ttl flag used by
174 # keepstore servers. Otherwise, reading data blocks and saving
175 # collections will fail with HTTP 403 permission errors.
177 # Modifying blob_signature_ttl invalidates existing signatures; see
178 # blob_signing_key note above.
180 # The default is 2 weeks.
181 blob_signature_ttl: 1209600
183 # Default lifetime for ephemeral collections: 2 weeks. This must not
184 # be less than blob_signature_ttl.
185 default_trash_lifetime: 1209600
187 # Interval (seconds) between trash sweeps. During a trash sweep,
188 # collections are marked as trash if their trash_at time has
189 # arrived, and deleted if their delete_at time has arrived.
190 trash_sweep_interval: 60
192 # Maximum characters of (JSON-encoded) query parameters to include
193 # in each request log entry. When params exceed this size, they will
194 # be JSON-encoded, truncated to this size, and logged as
196 max_request_log_params_size: 2000
198 # Maximum size (in bytes) allowed for a single API request. This
199 # limit is published in the discovery document for use by clients.
200 # Note: You must separately configure the upstream web server or
201 # proxy to actually enforce the desired maximum request size on the
203 max_request_size: 134217728
205 # Limit the number of bytes read from the database during an index
206 # request (by retrieving and returning fewer rows than would
207 # normally be returned in a single response).
208 # Note 1: This setting never reduces the number of returned rows to
209 # zero, no matter how big the first data row is.
210 # Note 2: Currently, this is only checked against a specific set of
211 # columns that tend to get large (collections.manifest_text,
212 # containers.mounts, workflows.definition). Other fields (e.g.,
213 # "properties" hashes) are not counted against this limit.
214 max_index_database_read: 134217728
216 # Maximum number of items to return when responding to a APIs that
217 # can return partial result sets using limit and offset parameters
218 # (e.g., *.index, groups.contents). If a request specifies a "limit"
219 # parameter higher than this value, this value is used instead.
220 max_items_per_response: 1000
222 # When you run the db:delete_old_job_logs task, it will find jobs that
223 # have been finished for at least this many seconds, and delete their
224 # stderr logs from the logs table.
225 clean_job_log_rows_after: <%= 30.days %>
227 # When you run the db:delete_old_container_logs task, it will find
228 # containers that have been finished for at least this many seconds,
229 # and delete their stdout, stderr, arv-mount, crunch-run, and
230 # crunchstat logs from the logs table.
231 clean_container_log_rows_after: <%= 30.days %>
233 # Time to keep audit logs, in seconds. (An audit log is a row added
234 # to the "logs" table in the PostgreSQL database each time an
235 # Arvados object is created, modified, or deleted.)
237 # Currently, websocket event notifications rely on audit logs, so
238 # this should not be set lower than 600 (5 minutes).
239 max_audit_log_age: 1209600
241 # Maximum number of log rows to delete in a single SQL transaction.
243 # If max_audit_log_delete_batch is 0, log entries will never be
244 # deleted by Arvados. Cleanup can be done by an external process
245 # without affecting any Arvados system processes, as long as very
246 # recent (<5 minutes old) logs are not deleted.
248 # 100000 is a reasonable batch size for most sites.
249 max_audit_log_delete_batch: 0
251 # The maximum number of compute nodes that can be in use simultaneously
252 # If this limit is reduced, any existing nodes with slot number >= new limit
253 # will not be counted against the new limit. In other words, the new limit
254 # won't be strictly enforced until those nodes with higher slot numbers
256 max_compute_nodes: 64
258 # These two settings control how frequently log events are flushed to the
259 # database. Log lines are buffered until either crunch_log_bytes_per_event
260 # has been reached or crunch_log_seconds_between_events has elapsed since
262 crunch_log_bytes_per_event: 4096
263 crunch_log_seconds_between_events: 1
265 # The sample period for throttling logs, in seconds.
266 crunch_log_throttle_period: 60
268 # Maximum number of bytes that job can log over crunch_log_throttle_period
269 # before being silenced until the end of the period.
270 crunch_log_throttle_bytes: 65536
272 # Maximum number of lines that job can log over crunch_log_throttle_period
273 # before being silenced until the end of the period.
274 crunch_log_throttle_lines: 1024
276 # Maximum bytes that may be logged by a single job. Log bytes that are
277 # silenced by throttling are not counted against this total.
278 crunch_limit_log_bytes_per_job: 67108864
280 crunch_log_partial_line_throttle_period: 5
282 # Attributes to suppress in events and audit logs. Notably,
283 # specifying ["manifest_text"] here typically makes the database
284 # smaller and faster.
286 # Warning: Using any non-empty value here can have undesirable side
287 # effects for any client or component that relies on event logs.
288 # Use at your own risk.
289 unlogged_attributes: []
291 # API methods to disable. Disabled methods are not listed in the
292 # discovery document, and respond 404 to all requests.
293 # Example: ["jobs.create", "pipeline_instances.create"]
294 disable_api_methods: []
297 ### Crunch, DNS & compute node management
300 # Preemptible instance support (e.g. AWS Spot Instances)
301 # When true, child containers will get created with the preemptible
302 # scheduling parameter parameter set.
303 preemptible_instances: false
305 # Docker image to be used when none found in runtime_constraints of a job
306 default_docker_image_for_jobs: false
308 # List of supported Docker Registry image formats that compute nodes
309 # are able to use. `arv keep docker` will error out if a user tries
310 # to store an image with an unsupported format. Use an empty array
311 # to skip the compatibility check (and display a warning message to
314 # Example for sites running docker < 1.10: ["v1"]
315 # Example for sites running docker >= 1.10: ["v2"]
316 # Example for disabling check: []
317 docker_image_formats: ["v2"]
319 # :none or :slurm_immediate
320 crunch_job_wrapper: :none
322 # username, or false = do not set uid when running jobs.
323 crunch_job_user: crunch
325 # The web service must be able to create/write this file, and
326 # crunch-job must be able to stat() it.
327 crunch_refresh_trigger: /tmp/crunch_refresh_trigger
329 # Path to dns server configuration directory
330 # (e.g. /etc/unbound.d/conf.d). If false, do not write any config
331 # files or touch restart.txt (see below).
332 dns_server_conf_dir: false
334 # Template file for the dns server host snippets. See
335 # unbound.template in this directory for an example. If false, do
336 # not write any config files.
337 dns_server_conf_template: false
339 # String to write to {dns_server_conf_dir}/restart.txt (with a
340 # trailing newline) after updating local data. If false, do not
341 # open or write the restart.txt file.
342 dns_server_reload_command: false
344 # Command to run after each DNS update. Template variables will be
345 # substituted; see the "unbound" example below. If false, do not run
347 dns_server_update_command: false
349 ## Example for unbound:
350 #dns_server_conf_dir: /etc/unbound/conf.d
351 #dns_server_conf_template: /path/to/your/api/server/config/unbound.template
352 ## ...plus one of the following two methods of reloading:
353 #dns_server_reload_command: unbound-control reload
354 #dns_server_update_command: echo %{hostname} %{hostname}.%{uuid_prefix} %{hostname}.%{uuid_prefix}.arvadosapi.com %{ptr_domain} | xargs -n 1 unbound-control local_data_remove && unbound-control local_data %{hostname} IN A %{ip_address} && unbound-control local_data %{hostname}.%{uuid_prefix} IN A %{ip_address} && unbound-control local_data %{hostname}.%{uuid_prefix}.arvadosapi.com IN A %{ip_address} && unbound-control local_data %{ptr_domain}. IN PTR %{hostname}.%{uuid_prefix}.arvadosapi.com
356 compute_node_domain: false
357 compute_node_nameservers:
360 # Hostname to assign to a compute node when it sends a "ping" and the
361 # hostname in its Node record is nil.
362 # During bootstrapping, the "ping" script is expected to notice the
363 # hostname given in the ping response, and update its unix hostname
365 # If false, leave the hostname alone (this is appropriate if your compute
366 # nodes' hostnames are already assigned by some other mechanism).
368 # One way or another, the hostnames of your node records should agree
369 # with your DNS records and your /etc/slurm-llnl/slurm.conf files.
371 # Example for compute0000, compute0001, ....:
372 # assign_node_hostname: compute%<slot_number>04d
373 # (See http://ruby-doc.org/core-2.2.2/Kernel.html#method-i-format for more.)
374 assign_node_hostname: compute%<slot_number>d
378 ### Job and container reuse logic.
381 # Include details about job reuse decisions in the server log. This
382 # causes additional database queries to run, so it should not be
383 # enabled unless you expect to examine the resulting logs for
384 # troubleshooting purposes.
385 log_reuse_decisions: false
387 # Control job reuse behavior when two completed jobs match the
388 # search criteria and have different outputs.
390 # If true, in case of a conflict, reuse the earliest job (this is
391 # similar to container reuse behavior).
393 # If false, in case of a conflict, do not reuse any completed job,
394 # but do reuse an already-running job if available (this is the
395 # original job reuse behavior, and is still the default).
396 reuse_job_if_outputs_differ: false
399 ### Federation support.
402 # You can enable use of this cluster by users who are authenticated
403 # by a remote Arvados site. Control which remote hosts are trusted
404 # to authenticate which user IDs by configuring remote_hosts,
405 # remote_hosts_via_dns, or both. The default configuration disables
406 # remote authentication.
408 # Map known prefixes to hosts. For example, if user IDs beginning
409 # with "zzzzz-" should be authenticated by the Arvados server at
410 # "zzzzz.example.com", use:
413 # zzzzz: zzzzz.example.com
416 # Use {prefix}.arvadosapi.com for any prefix not given in
417 # remote_hosts above.
418 remote_hosts_via_dns: false
420 # List of cluster prefixes. These are "trusted" clusters, users
421 # from the clusters listed here will be automatically setup and
422 # activated. This is separate from the settings
423 # auto_setup_new_users and new_users_are_active.
424 auto_activate_users_from: []
427 ### Remaining assorted configuration options.
430 arvados_theme: default
432 # Permit insecure (OpenSSL::SSL::VERIFY_NONE) connections to the
433 # Single Sign On (sso) server and remote Arvados sites. Should only
434 # be enabled during development when the SSO server is using a
438 ## Set Time.zone default to the specified zone and make Active
439 ## Record auto-convert to this zone. Run "rake -D time" for a list
440 ## of tasks for finding time zone names. Default is UTC.
441 #time_zone: Central Time (US & Canada)
443 ## Default encoding used in templates for Ruby 1.9.
446 # Enable the asset pipeline
449 # Version of your assets, change this if you want to expire all your assets
450 assets.version: "1.0"
452 # Allow clients to create collections by providing a manifest with
453 # unsigned data blob locators. IMPORTANT: This effectively disables
454 # access controls for data stored in Keep: a client who knows a hash
455 # can write a manifest that references the hash, pass it to
456 # collections.create (which will create a permission link), use
457 # collections.get to obtain a signature for that data locator, and
458 # use that signed locator to retrieve the data from Keep. Therefore,
459 # do not turn this on if your users expect to keep data private from
461 permit_create_collection_with_unsigned_manifest: false
463 default_openid_prefix: https://www.google.com/accounts/o8/id
465 # Override the automatic version string. With the default value of
466 # false, the version string is read from git-commit.version in
467 # Rails.root (included in vendor packages) or determined by invoking
469 source_version: false
471 # Override the automatic package version string. With the default version of
472 # false, the package version is read from package-build.version in Rails.root
473 # (included in vendor packages).
474 package_version: false
476 # Enable asynchronous permission graph rebuild. Must run
477 # script/permission-updater.rb as a separate process. When the permission
478 # cache is invalidated, the background process will update the permission
479 # graph cache. This feature is experimental!
480 async_permissions_update: false
482 # Default value for container_count_max for container requests. This is the
483 # number of times Arvados will create a new container to satisfy a container
484 # request. If a container is cancelled it will retry a new container if
485 # container_count < container_count_max on any container requests associated
486 # with the cancelled container.
487 container_count_max: 3
489 # Default value for keep_cache_ram of a container's runtime_constraints.
490 container_default_keep_cache_ram: 268435456
492 # Token to be included in all healthcheck requests. Disabled by default.
493 # Server expects request header of the format "Authorization: Bearer xxx"
494 ManagementToken: false
496 # URL of keep-web service. Provides read/write access to collections via
497 # HTTP and WebDAV protocols.
500 # keep_web_service_url: https://download.uuid_prefix.arvadosapi.com/
501 keep_web_service_url: false
507 consider_all_requests_local: true
508 action_controller.perform_caching: false
509 action_mailer.raise_delivery_errors: false
510 action_mailer.perform_deliveries: false
511 active_support.deprecation: :log
512 action_dispatch.best_standards_support: :builtin
513 active_record.auto_explain_threshold_in_seconds: 0.5
514 assets.compress: false
520 consider_all_requests_local: false
521 action_controller.perform_caching: true
522 serve_static_files: false
523 assets.compress: true
524 assets.compile: false
530 serve_static_files: true
531 static_cache_control: public, max-age=3600
533 consider_all_requests_local: true
534 action_controller.perform_caching: false
535 action_dispatch.show_exceptions: false
536 action_controller.allow_forgery_protection: false
537 action_mailer.delivery_method: :test
538 active_support.deprecation: :stderr
540 sso_app_id: arvados-server
541 sso_app_secret: <%= rand(2**512).to_s(36) %>
542 sso_provider_url: http://localhost:3002
543 secret_token: <%= rand(2**512).to_s(36) %>
544 blob_signing_key: zfhgfenhffzltr9dixws36j1yhksjoll2grmku38mi7yxd66h5j4q9w4jzanezacp8s6q0ro3hxakfye02152hncy6zml2ed0uc
545 user_profile_notification_address: arvados@example.com
546 workbench_address: https://localhost:3001/
547 git_repositories_dir: <%= Rails.root.join 'tmp', 'git', 'test' %>
548 git_internal_dir: <%= Rails.root.join 'tmp', 'internal.git' %>
549 websocket_address: "wss://0.0.0.0:<%= ENV['ARVADOS_TEST_WSS_PORT'] %>/websocket"
550 trash_sweep_interval: -1
551 docker_image_formats: ["v1"]