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
6 # /etc/arvados/config.yml instead.
8 # The order of precedence (highest to lowest):
9 # 1. Legacy component-specific config files (deprecated)
10 # 2. /etc/arvados/config.yml
11 # 3. config.default.yml
17 # Token to be included in all healthcheck requests. Disabled by default.
18 # Server expects request header of the format "Authorization: Bearer xxx"
63 # ShellInABox service endpoint URL for a given VM. If empty, do not
64 # offer web shell logins.
66 # E.g., using a path-based proxy server to forward connections to shell hosts:
67 # https://webshell.uuid_prefix.arvadosapi.com
69 # E.g., using a name-based proxy server to forward connections to shell hosts:
70 # https://*.webshell.uuid_prefix.arvadosapi.com
86 # max concurrent connections per arvados server daemon
89 # All parameters here are passed to the PG client library in a connection string;
90 # see https://www.postgresql.org/docs/current/static/libpq-connect.html#LIBPQ-PARAMKEYWORDS
97 # Maximum size (in bytes) allowed for a single API request. This
98 # limit is published in the discovery document for use by clients.
99 # Note: You must separately configure the upstream web server or
100 # proxy to actually enforce the desired maximum request size on the
102 MaxRequestSize: 134217728
104 # Limit the number of bytes read from the database during an index
105 # request (by retrieving and returning fewer rows than would
106 # normally be returned in a single response).
107 # Note 1: This setting never reduces the number of returned rows to
108 # zero, no matter how big the first data row is.
109 # Note 2: Currently, this is only checked against a specific set of
110 # columns that tend to get large (collections.manifest_text,
111 # containers.mounts, workflows.definition). Other fields (e.g.,
112 # "properties" hashes) are not counted against this limit.
113 MaxIndexDatabaseRead: 134217728
115 # Maximum number of items to return when responding to a APIs that
116 # can return partial result sets using limit and offset parameters
117 # (e.g., *.index, groups.contents). If a request specifies a "limit"
118 # parameter higher than this value, this value is used instead.
119 MaxItemsPerResponse: 1000
121 # API methods to disable. Disabled methods are not listed in the
122 # discovery document, and respond 404 to all requests.
123 # Example: ["jobs.create", "pipeline_instances.create"]
126 # Interval (seconds) between asynchronous permission view updates. Any
127 # permission-updating API called with the 'async' parameter schedules a an
128 # update on the permission view in the future, if not already scheduled.
129 AsyncPermissionsUpdateInterval: 20s
131 # Maximum number of concurrent outgoing requests to make while
132 # serving a single incoming multi-cluster (federated) request.
133 MaxRequestAmplification: 4
135 # RailsSessionSecretToken is a string of alphanumeric characters
136 # used by Rails to sign session tokens. IMPORTANT: This is a
137 # site secret. It should be at least 50 characters.
138 RailsSessionSecretToken: ""
140 # Maximum wall clock time to spend handling an incoming request.
144 # Config parameters to automatically setup new users. If enabled,
145 # this users will be able to self-activate. Enable this if you want
146 # to run an open instance where anyone can create an account and use
147 # the system without requiring manual approval.
149 # The params auto_setup_new_users_with_* are meaningful only when auto_setup_new_users is turned on.
150 # auto_setup_name_blacklist is a list of usernames to be blacklisted for auto setup.
151 AutoSetupNewUsers: false
152 AutoSetupNewUsersWithVmUUID: ""
153 AutoSetupNewUsersWithRepository: false
154 AutoSetupUsernameBlacklist: [arvados, git, gitolite, gitolite-admin, root, syslog]
156 # When new_users_are_active is set to true, new users will be active
157 # immediately. This skips the "self-activate" step which enforces
158 # user agreements. Should only be enabled for development.
159 NewUsersAreActive: false
161 # The e-mail address of the user you would like to become marked as an admin
162 # user on their first login.
163 # In the default configuration, authentication happens through the Arvados SSO
164 # server, which uses OAuth2 against Google's servers, so in that case this
165 # should be an address associated with a Google account.
166 AutoAdminUserWithEmail: ""
168 # If auto_admin_first_user is set to true, the first user to log in when no
169 # other admin users exist will automatically become an admin user.
170 AutoAdminFirstUser: false
172 # Email address to notify whenever a user creates a profile for the
174 UserProfileNotificationAddress: ""
175 AdminNotifierEmailFrom: arvados@example.com
176 EmailSubjectPrefix: "[ARVADOS] "
177 UserNotifierEmailFrom: arvados@example.com
178 NewUserNotificationRecipients: []
179 NewInactiveUserNotificationRecipients: []
181 # Set anonymous_user_token to enable anonymous user access. You can get
182 # the token by running "bundle exec ./script/get_anonymous_user_token.rb"
183 # in the directory where your API server is running.
184 AnonymousUserToken: ""
187 # Time to keep audit logs, in seconds. (An audit log is a row added
188 # to the "logs" table in the PostgreSQL database each time an
189 # Arvados object is created, modified, or deleted.)
191 # Currently, websocket event notifications rely on audit logs, so
192 # this should not be set lower than 300 (5 minutes).
195 # Maximum number of log rows to delete in a single SQL transaction.
197 # If max_audit_log_delete_batch is 0, log entries will never be
198 # deleted by Arvados. Cleanup can be done by an external process
199 # without affecting any Arvados system processes, as long as very
200 # recent (<5 minutes old) logs are not deleted.
202 # 100000 is a reasonable batch size for most sites.
205 # Attributes to suppress in events and audit logs. Notably,
206 # specifying ["manifest_text"] here typically makes the database
207 # smaller and faster.
209 # Warning: Using any non-empty value here can have undesirable side
210 # effects for any client or component that relies on event logs.
211 # Use at your own risk.
212 UnloggedAttributes: []
216 # Logging threshold: panic, fatal, error, warn, info, debug, or
220 # Logging format: json or text
223 # Maximum characters of (JSON-encoded) query parameters to include
224 # in each request log entry. When params exceed this size, they will
225 # be JSON-encoded, truncated to this size, and logged as
227 MaxRequestLogParamsSize: 2000
230 # Allow clients to create collections by providing a manifest with
231 # unsigned data blob locators. IMPORTANT: This effectively disables
232 # access controls for data stored in Keep: a client who knows a hash
233 # can write a manifest that references the hash, pass it to
234 # collections.create (which will create a permission link), use
235 # collections.get to obtain a signature for that data locator, and
236 # use that signed locator to retrieve the data from Keep. Therefore,
237 # do not turn this on if your users expect to keep data private from
241 # blob_signing_key is a string of alphanumeric characters used to
242 # generate permission signatures for Keep locators. It must be
243 # identical to the permission key given to Keep. IMPORTANT: This is
244 # a site secret. It should be at least 50 characters.
246 # Modifying blob_signing_key will invalidate all existing
247 # signatures, which can cause programs to fail (e.g., arv-put,
248 # arv-get, and Crunch jobs). To avoid errors, rotate keys only when
249 # no such processes are running.
252 # Default replication level for collections. This is used when a
253 # collection's replication_desired attribute is nil.
254 DefaultReplication: 2
256 # Lifetime (in seconds) of blob permission signatures generated by
257 # the API server. This determines how long a client can take (after
258 # retrieving a collection record) to retrieve the collection data
259 # from Keep. If the client needs more time than that (assuming the
260 # collection still has the same content and the relevant user/token
261 # still has permission) the client can retrieve the collection again
262 # to get fresh signatures.
264 # This must be exactly equal to the -blob-signature-ttl flag used by
265 # keepstore servers. Otherwise, reading data blocks and saving
266 # collections will fail with HTTP 403 permission errors.
268 # Modifying blob_signature_ttl invalidates existing signatures; see
269 # blob_signing_key note above.
271 # The default is 2 weeks.
274 # Default lifetime for ephemeral collections: 2 weeks. This must not
275 # be less than blob_signature_ttl.
276 DefaultTrashLifetime: 336h
278 # Interval (seconds) between trash sweeps. During a trash sweep,
279 # collections are marked as trash if their trash_at time has
280 # arrived, and deleted if their delete_at time has arrived.
281 TrashSweepInterval: 60s
283 # If true, enable collection versioning.
284 # When a collection's preserve_version field is true or the current version
285 # is older than the amount of seconds defined on preserve_version_if_idle,
286 # a snapshot of the collection's previous state is created and linked to
287 # the current collection.
288 CollectionVersioning: false
290 # 0s = auto-create a new version on every update.
291 # -1s = never auto-create new versions.
292 # > 0s = auto-create a new version when older than the specified number of seconds.
293 PreserveVersionIfIdle: -1s
295 # Managed collection properties. At creation time, if the client didn't
296 # provide the listed keys, they will be automatically populated following
297 # one of the following behaviors:
299 # * UUID of the user who owns the containing project.
300 # responsible_person_uuid: {Function: original_owner, Protected: true}
302 # * Default concrete value.
303 # foo_bar: {Value: baz, Protected: false}
305 # If Protected is true, only an admin user can modify its value.
307 SAMPLE: {Function: original_owner, Protected: true}
310 # These settings are provided by your OAuth2 provider (e.g.,
312 ProviderAppSecret: ""
316 # Git repositories must be readable by api server, or you won't be
317 # able to submit crunch jobs. To pass the test suites, put a clone
318 # of the arvados tree in {git_repositories_dir}/arvados.git or
319 # {git_repositories_dir}/arvados/.git
320 Repositories: /var/lib/arvados/git/repositories
328 # List of supported Docker Registry image formats that compute nodes
329 # are able to use. `arv keep docker` will error out if a user tries
330 # to store an image with an unsupported format. Use an empty array
331 # to skip the compatibility check (and display a warning message to
334 # Example for sites running docker < 1.10: ["v1"]
335 # Example for sites running docker >= 1.10: ["v2"]
336 # Example for disabling check: []
337 SupportedDockerImageFormats: ["v2"]
339 # Include details about job reuse decisions in the server log. This
340 # causes additional database queries to run, so it should not be
341 # enabled unless you expect to examine the resulting logs for
342 # troubleshooting purposes.
343 LogReuseDecisions: false
345 # Default value for keep_cache_ram of a container's runtime_constraints.
346 DefaultKeepCacheRAM: 268435456
348 # Number of times a container can be unlocked before being
349 # automatically cancelled.
350 MaxDispatchAttempts: 5
352 # Default value for container_count_max for container requests. This is the
353 # number of times Arvados will create a new container to satisfy a container
354 # request. If a container is cancelled it will retry a new container if
355 # container_count < container_count_max on any container requests associated
356 # with the cancelled container.
359 # The maximum number of compute nodes that can be in use simultaneously
360 # If this limit is reduced, any existing nodes with slot number >= new limit
361 # will not be counted against the new limit. In other words, the new limit
362 # won't be strictly enforced until those nodes with higher slot numbers
366 # Preemptible instance support (e.g. AWS Spot Instances)
367 # When true, child containers will get created with the preemptible
368 # scheduling parameter parameter set.
369 UsePreemptibleInstances: false
371 # PEM encoded SSH key (RSA, DSA, or ECDSA) used by the
372 # (experimental) cloud dispatcher for executing containers on
373 # worker VMs. Begins with "-----BEGIN RSA PRIVATE KEY-----\n"
374 # and ends with "\n-----END RSA PRIVATE KEY-----\n".
375 DispatchPrivateKey: none
377 # Maximum time to wait for workers to come up before abandoning
378 # stale locks from a previous dispatch process.
382 # When you run the db:delete_old_container_logs task, it will find
383 # containers that have been finished for at least this many seconds,
384 # and delete their stdout, stderr, arv-mount, crunch-run, and
385 # crunchstat logs from the logs table.
388 # These two settings control how frequently log events are flushed to the
389 # database. Log lines are buffered until either crunch_log_bytes_per_event
390 # has been reached or crunch_log_seconds_between_events has elapsed since
392 LogBytesPerEvent: 4096
393 LogSecondsBetweenEvents: 1
395 # The sample period for throttling logs.
396 LogThrottlePeriod: 60s
398 # Maximum number of bytes that job can log over crunch_log_throttle_period
399 # before being silenced until the end of the period.
400 LogThrottleBytes: 65536
402 # Maximum number of lines that job can log over crunch_log_throttle_period
403 # before being silenced until the end of the period.
404 LogThrottleLines: 1024
406 # Maximum bytes that may be logged by a single job. Log bytes that are
407 # silenced by throttling are not counted against this total.
408 LimitLogBytesPerJob: 67108864
410 LogPartialLineThrottlePeriod: 5s
412 # Container logs are written to Keep and saved in a
413 # collection, which is updated periodically while the
414 # container runs. This value sets the interval between
415 # collection updates.
418 # The log collection is also updated when the specified amount of
419 # log data (given in bytes) is produced in less than one update
425 # Path to dns server configuration directory
426 # (e.g. /etc/unbound.d/conf.d). If false, do not write any config
427 # files or touch restart.txt (see below).
430 # Template file for the dns server host snippets. See
431 # unbound.template in this directory for an example. If false, do
432 # not write any config files.
433 DNSServerConfTemplate: ""
435 # String to write to {dns_server_conf_dir}/restart.txt (with a
436 # trailing newline) after updating local data. If false, do not
437 # open or write the restart.txt file.
438 DNSServerReloadCommand: ""
440 # Command to run after each DNS update. Template variables will be
441 # substituted; see the "unbound" example below. If false, do not run
443 DNSServerUpdateCommand: ""
445 ComputeNodeDomain: ""
446 ComputeNodeNameservers:
449 # Hostname to assign to a compute node when it sends a "ping" and the
450 # hostname in its Node record is nil.
451 # During bootstrapping, the "ping" script is expected to notice the
452 # hostname given in the ping response, and update its unix hostname
454 # If false, leave the hostname alone (this is appropriate if your compute
455 # nodes' hostnames are already assigned by some other mechanism).
457 # One way or another, the hostnames of your node records should agree
458 # with your DNS records and your /etc/slurm-llnl/slurm.conf files.
460 # Example for compute0000, compute0001, ....:
461 # assign_node_hostname: compute%<slot_number>04d
462 # (See http://ruby-doc.org/core-2.2.2/Kernel.html#method-i-format for more.)
463 AssignNodeHostname: "compute%<slot_number>d"
466 # Enable the legacy Jobs API. This value must be a string.
467 # 'auto' -- (default) enable the Jobs API only if it has been used before
468 # (i.e., there are job records in the database)
469 # 'true' -- enable the Jobs API despite lack of existing records.
470 # 'false' -- disable the Jobs API despite presence of existing records.
473 # Git repositories must be readable by api server, or you won't be
474 # able to submit crunch jobs. To pass the test suites, put a clone
475 # of the arvados tree in {git_repositories_dir}/arvados.git or
476 # {git_repositories_dir}/arvados/.git
477 GitInternalDir: /var/lib/arvados/internal.git
479 # Docker image to be used when none found in runtime_constraints of a job
480 DefaultDockerImage: ""
482 # none or slurm_immediate
483 CrunchJobWrapper: none
485 # username, or false = do not set uid when running jobs.
486 CrunchJobUser: crunch
488 # The web service must be able to create/write this file, and
489 # crunch-job must be able to stat() it.
490 CrunchRefreshTrigger: /tmp/crunch_refresh_trigger
492 # Control job reuse behavior when two completed jobs match the
493 # search criteria and have different outputs.
495 # If true, in case of a conflict, reuse the earliest job (this is
496 # similar to container reuse behavior).
498 # If false, in case of a conflict, do not reuse any completed job,
499 # but do reuse an already-running job if available (this is the
500 # original job reuse behavior, and is still the default).
501 ReuseJobIfOutputsDiffer: false
504 # Enable the cloud scheduler (experimental).
507 # Name/number of port where workers' SSH services listen.
510 # Interval between queue polls.
513 # Shell command to execute on each worker to determine whether
514 # the worker is booted and ready to run containers. It should
515 # exit zero if the worker is ready.
516 BootProbeCommand: "docker ps -q"
518 # Minimum interval between consecutive probes to a single
522 # Maximum probes per second, across all workers in a pool.
523 MaxProbesPerSecond: 10
525 # Time before repeating SIGTERM when killing a container.
528 # Time to give up on SIGTERM and write off the worker.
531 # Maximum create/destroy-instance operations per second (0 =
533 MaxCloudOpsPerSecond: 0
535 # Interval between cloud provider syncs/updates ("list all
539 # Time to leave an idle worker running (in case new containers
540 # appear in the queue that it can run) before shutting it
544 # Time to wait for a new worker to boot (i.e., pass
545 # BootProbeCommand) before giving up and shutting it down.
548 # Maximum time a worker can stay alive with no successful
549 # probes before being automatically shut down.
552 # Time after shutting down a worker to retry the
553 # shutdown/destroy operation.
556 # Worker VM image ID.
559 # Tags to add on all resources (VMs, NICs, disks) created by
560 # the container dispatcher. (Arvados's own tags --
561 # InstanceType, IdleBehavior, and InstanceSecret -- will also
566 # Prefix for predefined tags used by Arvados (InstanceSetID,
567 # InstanceType, InstanceSecret, IdleBehavior). With the
568 # default value "Arvados", tags are "ArvadosInstanceSetID",
569 # "ArvadosInstanceSecret", etc.
571 # This should only be changed while no cloud resources are in
572 # use and the cloud dispatcher is not running. Otherwise,
573 # VMs/resources that were added using the old tag prefix will
574 # need to be detected and cleaned up manually.
575 TagKeyPrefix: Arvados
577 # Cloud driver: "azure" (Microsoft Azure) or "ec2" (Amazon AWS).
580 # Cloud-specific driver parameters.
587 # (ec2) Instance configuration.
593 AdminUsername: debian
595 # (azure) Credentials.
601 # (azure) Instance configuration.
602 CloudEnvironment: AzurePublicCloud
609 DeleteDanglingResourcesAfter: 20s
610 AdminUsername: arvados
614 # Use the instance type name as the key (in place of "SAMPLE" in
615 # this sample entry).
617 # Cloud provider's instance type. Defaults to the configured type name.
621 IncludedScratch: 16GB
629 SendUserSetupNotificationEmail: true
630 IssueReporterEmailFrom: ""
631 IssueReporterEmailTo: ""
632 SupportEmailAddress: ""
642 # API endpoint host or host:port; default is {id}.arvadosapi.com
643 Host: sample.arvadosapi.com
645 # Perform a proxy request when a local client requests an
646 # object belonging to this remote.
649 # Default "https". Can be set to "http" for testing.
652 # Disable TLS verify. Can be set to true for testing.
655 # When users present tokens issued by this remote cluster, and
656 # their accounts are active on the remote cluster, activate
657 # them on this cluster too.
663 ActivationContactLink: mailto:info@arvados.org
664 ArvadosDocsite: https://doc.arvados.org
665 ArvadosPublicDataDocURL: https://playground.arvados.org/projects/public
666 ShowUserAgreementInline: false
668 RepositoryCache: /var/www/arvados-workbench/current/tmp/git
669 UserProfileFormFields: {}
670 # exampleTextValue: # key that will be set in properties
673 # FormFieldDescription: ""
676 # exampleOptionsValue:
679 # FormFieldDescription: ""
686 UserProfileFormMessage: 'Welcome to Arvados. All <span style="color:red">required fields</span> must be completed before you can proceed.'
688 # Mimetypes of applications for which the view icon
689 # would be enabled in a collection's show page.
690 # It is sufficient to list only applications here.
691 # No need to list text and image types.
692 ApplicationMimetypesWithViewIcon:
708 LogViewerMaxBytes: 1M
709 EnablePublicProjectsPage: true
710 EnableGettingStartedPopup: false
711 APIResponseCompression: true
712 APIClientConnectTimeout: 2m
713 APIClientReceiveTimeout: 5m
714 RunningJobLogRecordsToFetch: 2000
715 ShowRecentCollectionsOnDashboard: true
716 ShowUserNotifications: true
719 SiteName: Arvados Workbench
720 TrustAllContent: false
724 FileViewersConfigURL: ""
726 # Use experimental controller code (see https://dev.arvados.org/issues/14287)
727 EnableBetaController14287: false