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"
51 # Base URL for Workbench inline preview. If blank, use
52 # WebDAVDownload instead, and disable inline preview.
53 # If both are empty, downloading collections from workbench
56 # It is important to properly configure the download service
57 # to migitate cross-site-scripting (XSS) attacks. A HTML page
58 # can be stored in collection. If an attacker causes a victim
59 # to visit that page through Workbench, it will be rendered by
60 # the browser. If all collections are served at the same
61 # domain, the browser will consider collections as coming from
62 # the same origin and having access to the same browsing data,
63 # enabling malicious Javascript on that page to access Arvados
64 # on behalf of the victim.
66 # This is mitigating by having separate domains for each
67 # collection, or limiting preview to circumstances where the
68 # collection is not accessed with the user's regular
71 # Serve preview links using uuid or pdh in subdomain
72 # (requires wildcard DNS and TLS certificate)
73 # https://*.collections.uuid_prefix.arvadosapi.com
75 # Serve preview links using uuid or pdh in main domain
76 # (requires wildcard DNS and TLS certificate)
77 # https://*--collections.uuid_prefix.arvadosapi.com
79 # Serve preview links by setting uuid or pdh in the path.
80 # This configuration only allows previews of public data or
81 # collection-sharing links, because these use the anonymous
82 # user token or the token is already embedded in the URL.
83 # Other data must be handled as downloads via WebDAVDownload:
84 # https://collections.uuid_prefix.arvadosapi.com
90 # Base URL for download links. If blank, serve links to WebDAV
91 # with disposition=attachment query param. Unlike preview links,
92 # browsers do not render attachments, so there is no risk of XSS.
94 # If WebDAVDownload is blank, and WebDAV uses a
95 # single-origin form, then Workbench will show an error page
97 # Serve download links by setting uuid or pdh in the path:
98 # https://download.uuid_prefix.arvadosapi.com
110 # ShellInABox service endpoint URL for a given VM. If empty, do not
111 # offer web shell logins.
113 # E.g., using a path-based proxy server to forward connections to shell hosts:
114 # https://webshell.uuid_prefix.arvadosapi.com
116 # E.g., using a name-based proxy server to forward connections to shell hosts:
117 # https://*.webshell.uuid_prefix.arvadosapi.com
133 # max concurrent connections per arvados server daemon
136 # All parameters here are passed to the PG client library in a connection string;
137 # see https://www.postgresql.org/docs/current/static/libpq-connect.html#LIBPQ-PARAMKEYWORDS
144 # Maximum size (in bytes) allowed for a single API request. This
145 # limit is published in the discovery document for use by clients.
146 # Note: You must separately configure the upstream web server or
147 # proxy to actually enforce the desired maximum request size on the
149 MaxRequestSize: 134217728
151 # Limit the number of bytes read from the database during an index
152 # request (by retrieving and returning fewer rows than would
153 # normally be returned in a single response).
154 # Note 1: This setting never reduces the number of returned rows to
155 # zero, no matter how big the first data row is.
156 # Note 2: Currently, this is only checked against a specific set of
157 # columns that tend to get large (collections.manifest_text,
158 # containers.mounts, workflows.definition). Other fields (e.g.,
159 # "properties" hashes) are not counted against this limit.
160 MaxIndexDatabaseRead: 134217728
162 # Maximum number of items to return when responding to a APIs that
163 # can return partial result sets using limit and offset parameters
164 # (e.g., *.index, groups.contents). If a request specifies a "limit"
165 # parameter higher than this value, this value is used instead.
166 MaxItemsPerResponse: 1000
168 # API methods to disable. Disabled methods are not listed in the
169 # discovery document, and respond 404 to all requests.
170 # Example: ["jobs.create", "pipeline_instances.create"]
173 # Interval (seconds) between asynchronous permission view updates. Any
174 # permission-updating API called with the 'async' parameter schedules a an
175 # update on the permission view in the future, if not already scheduled.
176 AsyncPermissionsUpdateInterval: 20s
178 # Maximum number of concurrent outgoing requests to make while
179 # serving a single incoming multi-cluster (federated) request.
180 MaxRequestAmplification: 4
182 # RailsSessionSecretToken is a string of alphanumeric characters
183 # used by Rails to sign session tokens. IMPORTANT: This is a
184 # site secret. It should be at least 50 characters.
185 RailsSessionSecretToken: ""
187 # Maximum wall clock time to spend handling an incoming request.
191 # Config parameters to automatically setup new users. If enabled,
192 # this users will be able to self-activate. Enable this if you want
193 # to run an open instance where anyone can create an account and use
194 # the system without requiring manual approval.
196 # The params auto_setup_new_users_with_* are meaningful only when auto_setup_new_users is turned on.
197 # auto_setup_name_blacklist is a list of usernames to be blacklisted for auto setup.
198 AutoSetupNewUsers: false
199 AutoSetupNewUsersWithVmUUID: ""
200 AutoSetupNewUsersWithRepository: false
201 AutoSetupUsernameBlacklist: [arvados, git, gitolite, gitolite-admin, root, syslog]
203 # When new_users_are_active is set to true, new users will be active
204 # immediately. This skips the "self-activate" step which enforces
205 # user agreements. Should only be enabled for development.
206 NewUsersAreActive: false
208 # The e-mail address of the user you would like to become marked as an admin
209 # user on their first login.
210 # In the default configuration, authentication happens through the Arvados SSO
211 # server, which uses OAuth2 against Google's servers, so in that case this
212 # should be an address associated with a Google account.
213 AutoAdminUserWithEmail: ""
215 # If auto_admin_first_user is set to true, the first user to log in when no
216 # other admin users exist will automatically become an admin user.
217 AutoAdminFirstUser: false
219 # Email address to notify whenever a user creates a profile for the
221 UserProfileNotificationAddress: ""
222 AdminNotifierEmailFrom: arvados@example.com
223 EmailSubjectPrefix: "[ARVADOS] "
224 UserNotifierEmailFrom: arvados@example.com
225 NewUserNotificationRecipients: []
226 NewInactiveUserNotificationRecipients: []
228 # Set anonymous_user_token to enable anonymous user access. You can get
229 # the token by running "bundle exec ./script/get_anonymous_user_token.rb"
230 # in the directory where your API server is running.
231 AnonymousUserToken: ""
234 # Time to keep audit logs, in seconds. (An audit log is a row added
235 # to the "logs" table in the PostgreSQL database each time an
236 # Arvados object is created, modified, or deleted.)
238 # Currently, websocket event notifications rely on audit logs, so
239 # this should not be set lower than 300 (5 minutes).
242 # Maximum number of log rows to delete in a single SQL transaction.
244 # If max_audit_log_delete_batch is 0, log entries will never be
245 # deleted by Arvados. Cleanup can be done by an external process
246 # without affecting any Arvados system processes, as long as very
247 # recent (<5 minutes old) logs are not deleted.
249 # 100000 is a reasonable batch size for most sites.
252 # Attributes to suppress in events and audit logs. Notably,
253 # specifying ["manifest_text"] here typically makes the database
254 # smaller and faster.
256 # Warning: Using any non-empty value here can have undesirable side
257 # effects for any client or component that relies on event logs.
258 # Use at your own risk.
259 UnloggedAttributes: []
263 # Logging threshold: panic, fatal, error, warn, info, debug, or
267 # Logging format: json or text
270 # Maximum characters of (JSON-encoded) query parameters to include
271 # in each request log entry. When params exceed this size, they will
272 # be JSON-encoded, truncated to this size, and logged as
274 MaxRequestLogParamsSize: 2000
277 # Allow clients to create collections by providing a manifest with
278 # unsigned data blob locators. IMPORTANT: This effectively disables
279 # access controls for data stored in Keep: a client who knows a hash
280 # can write a manifest that references the hash, pass it to
281 # collections.create (which will create a permission link), use
282 # collections.get to obtain a signature for that data locator, and
283 # use that signed locator to retrieve the data from Keep. Therefore,
284 # do not turn this on if your users expect to keep data private from
288 # BlobSigningKey is a string of alphanumeric characters used to
289 # generate permission signatures for Keep locators. It must be
290 # identical to the permission key given to Keep. IMPORTANT: This is
291 # a site secret. It should be at least 50 characters.
293 # Modifying blob_signing_key will invalidate all existing
294 # signatures, which can cause programs to fail (e.g., arv-put,
295 # arv-get, and Crunch jobs). To avoid errors, rotate keys only when
296 # no such processes are running.
299 # Default replication level for collections. This is used when a
300 # collection's replication_desired attribute is nil.
301 DefaultReplication: 2
303 # Lifetime (in seconds) of blob permission signatures generated by
304 # the API server. This determines how long a client can take (after
305 # retrieving a collection record) to retrieve the collection data
306 # from Keep. If the client needs more time than that (assuming the
307 # collection still has the same content and the relevant user/token
308 # still has permission) the client can retrieve the collection again
309 # to get fresh signatures.
311 # This must be exactly equal to the -blob-signature-ttl flag used by
312 # keepstore servers. Otherwise, reading data blocks and saving
313 # collections will fail with HTTP 403 permission errors.
315 # Modifying blob_signature_ttl invalidates existing signatures; see
316 # blob_signing_key note above.
318 # The default is 2 weeks.
321 # Default lifetime for ephemeral collections: 2 weeks. This must not
322 # be less than blob_signature_ttl.
323 DefaultTrashLifetime: 336h
325 # Interval (seconds) between trash sweeps. During a trash sweep,
326 # collections are marked as trash if their trash_at time has
327 # arrived, and deleted if their delete_at time has arrived.
328 TrashSweepInterval: 60s
330 # If true, enable collection versioning.
331 # When a collection's preserve_version field is true or the current version
332 # is older than the amount of seconds defined on preserve_version_if_idle,
333 # a snapshot of the collection's previous state is created and linked to
334 # the current collection.
335 CollectionVersioning: false
337 # 0s = auto-create a new version on every update.
338 # -1s = never auto-create new versions.
339 # > 0s = auto-create a new version when older than the specified number of seconds.
340 PreserveVersionIfIdle: -1s
342 # Managed collection properties. At creation time, if the client didn't
343 # provide the listed keys, they will be automatically populated following
344 # one of the following behaviors:
346 # * UUID of the user who owns the containing project.
347 # responsible_person_uuid: {Function: original_owner, Protected: true}
349 # * Default concrete value.
350 # foo_bar: {Value: baz, Protected: false}
352 # If Protected is true, only an admin user can modify its value.
354 SAMPLE: {Function: original_owner, Protected: true}
356 # In "trust all content" mode, Workbench will redirect download
357 # requests to WebDAV preview link, even in the cases when
358 # WebDAV would have to expose XSS vulnerabilities in order to
359 # handle the redirect (see discussion on Services.WebDAV).
361 # This setting has no effect in the recommended configuration,
362 # where the WebDAV is configured to have a separate domain for
363 # every collection; in this case XSS protection is provided by
364 # browsers' same-origin policy.
366 # The default setting (false) is appropriate for a multi-user site.
367 TrustAllContent: false
370 # These settings are provided by your OAuth2 provider (e.g.,
372 ProviderAppSecret: ""
376 # Git repositories must be readable by api server, or you won't be
377 # able to submit crunch jobs. To pass the test suites, put a clone
378 # of the arvados tree in {git_repositories_dir}/arvados.git or
379 # {git_repositories_dir}/arvados/.git
380 Repositories: /var/lib/arvados/git/repositories
388 # List of supported Docker Registry image formats that compute nodes
389 # are able to use. `arv keep docker` will error out if a user tries
390 # to store an image with an unsupported format. Use an empty array
391 # to skip the compatibility check (and display a warning message to
394 # Example for sites running docker < 1.10: ["v1"]
395 # Example for sites running docker >= 1.10: ["v2"]
396 # Example for disabling check: []
397 SupportedDockerImageFormats: ["v2"]
399 # Include details about job reuse decisions in the server log. This
400 # causes additional database queries to run, so it should not be
401 # enabled unless you expect to examine the resulting logs for
402 # troubleshooting purposes.
403 LogReuseDecisions: false
405 # Default value for keep_cache_ram of a container's runtime_constraints.
406 DefaultKeepCacheRAM: 268435456
408 # Number of times a container can be unlocked before being
409 # automatically cancelled.
410 MaxDispatchAttempts: 5
412 # Default value for container_count_max for container requests. This is the
413 # number of times Arvados will create a new container to satisfy a container
414 # request. If a container is cancelled it will retry a new container if
415 # container_count < container_count_max on any container requests associated
416 # with the cancelled container.
419 # The maximum number of compute nodes that can be in use simultaneously
420 # If this limit is reduced, any existing nodes with slot number >= new limit
421 # will not be counted against the new limit. In other words, the new limit
422 # won't be strictly enforced until those nodes with higher slot numbers
426 # Preemptible instance support (e.g. AWS Spot Instances)
427 # When true, child containers will get created with the preemptible
428 # scheduling parameter parameter set.
429 UsePreemptibleInstances: false
431 # PEM encoded SSH key (RSA, DSA, or ECDSA) used by the
432 # (experimental) cloud dispatcher for executing containers on
433 # worker VMs. Begins with "-----BEGIN RSA PRIVATE KEY-----\n"
434 # and ends with "\n-----END RSA PRIVATE KEY-----\n".
435 DispatchPrivateKey: none
437 # Maximum time to wait for workers to come up before abandoning
438 # stale locks from a previous dispatch process.
442 # When you run the db:delete_old_container_logs task, it will find
443 # containers that have been finished for at least this many seconds,
444 # and delete their stdout, stderr, arv-mount, crunch-run, and
445 # crunchstat logs from the logs table.
448 # These two settings control how frequently log events are flushed to the
449 # database. Log lines are buffered until either crunch_log_bytes_per_event
450 # has been reached or crunch_log_seconds_between_events has elapsed since
452 LogBytesPerEvent: 4096
453 LogSecondsBetweenEvents: 1
455 # The sample period for throttling logs.
456 LogThrottlePeriod: 60s
458 # Maximum number of bytes that job can log over crunch_log_throttle_period
459 # before being silenced until the end of the period.
460 LogThrottleBytes: 65536
462 # Maximum number of lines that job can log over crunch_log_throttle_period
463 # before being silenced until the end of the period.
464 LogThrottleLines: 1024
466 # Maximum bytes that may be logged by a single job. Log bytes that are
467 # silenced by throttling are not counted against this total.
468 LimitLogBytesPerJob: 67108864
470 LogPartialLineThrottlePeriod: 5s
472 # Container logs are written to Keep and saved in a
473 # collection, which is updated periodically while the
474 # container runs. This value sets the interval between
475 # collection updates.
478 # The log collection is also updated when the specified amount of
479 # log data (given in bytes) is produced in less than one update
485 # Path to dns server configuration directory
486 # (e.g. /etc/unbound.d/conf.d). If false, do not write any config
487 # files or touch restart.txt (see below).
490 # Template file for the dns server host snippets. See
491 # unbound.template in this directory for an example. If false, do
492 # not write any config files.
493 DNSServerConfTemplate: ""
495 # String to write to {dns_server_conf_dir}/restart.txt (with a
496 # trailing newline) after updating local data. If false, do not
497 # open or write the restart.txt file.
498 DNSServerReloadCommand: ""
500 # Command to run after each DNS update. Template variables will be
501 # substituted; see the "unbound" example below. If false, do not run
503 DNSServerUpdateCommand: ""
505 ComputeNodeDomain: ""
506 ComputeNodeNameservers:
509 # Hostname to assign to a compute node when it sends a "ping" and the
510 # hostname in its Node record is nil.
511 # During bootstrapping, the "ping" script is expected to notice the
512 # hostname given in the ping response, and update its unix hostname
514 # If false, leave the hostname alone (this is appropriate if your compute
515 # nodes' hostnames are already assigned by some other mechanism).
517 # One way or another, the hostnames of your node records should agree
518 # with your DNS records and your /etc/slurm-llnl/slurm.conf files.
520 # Example for compute0000, compute0001, ....:
521 # assign_node_hostname: compute%<slot_number>04d
522 # (See http://ruby-doc.org/core-2.2.2/Kernel.html#method-i-format for more.)
523 AssignNodeHostname: "compute%<slot_number>d"
526 # Enable the legacy Jobs API. This value must be a string.
527 # 'auto' -- (default) enable the Jobs API only if it has been used before
528 # (i.e., there are job records in the database)
529 # 'true' -- enable the Jobs API despite lack of existing records.
530 # 'false' -- disable the Jobs API despite presence of existing records.
533 # Git repositories must be readable by api server, or you won't be
534 # able to submit crunch jobs. To pass the test suites, put a clone
535 # of the arvados tree in {git_repositories_dir}/arvados.git or
536 # {git_repositories_dir}/arvados/.git
537 GitInternalDir: /var/lib/arvados/internal.git
539 # Docker image to be used when none found in runtime_constraints of a job
540 DefaultDockerImage: ""
542 # none or slurm_immediate
543 CrunchJobWrapper: none
545 # username, or false = do not set uid when running jobs.
546 CrunchJobUser: crunch
548 # The web service must be able to create/write this file, and
549 # crunch-job must be able to stat() it.
550 CrunchRefreshTrigger: /tmp/crunch_refresh_trigger
552 # Control job reuse behavior when two completed jobs match the
553 # search criteria and have different outputs.
555 # If true, in case of a conflict, reuse the earliest job (this is
556 # similar to container reuse behavior).
558 # If false, in case of a conflict, do not reuse any completed job,
559 # but do reuse an already-running job if available (this is the
560 # original job reuse behavior, and is still the default).
561 ReuseJobIfOutputsDiffer: false
564 # Enable the cloud scheduler (experimental).
567 # Name/number of port where workers' SSH services listen.
570 # Interval between queue polls.
573 # Shell command to execute on each worker to determine whether
574 # the worker is booted and ready to run containers. It should
575 # exit zero if the worker is ready.
576 BootProbeCommand: "docker ps -q"
578 # Minimum interval between consecutive probes to a single
582 # Maximum probes per second, across all workers in a pool.
583 MaxProbesPerSecond: 10
585 # Time before repeating SIGTERM when killing a container.
588 # Time to give up on SIGTERM and write off the worker.
591 # Maximum create/destroy-instance operations per second (0 =
593 MaxCloudOpsPerSecond: 0
595 # Interval between cloud provider syncs/updates ("list all
599 # Time to leave an idle worker running (in case new containers
600 # appear in the queue that it can run) before shutting it
604 # Time to wait for a new worker to boot (i.e., pass
605 # BootProbeCommand) before giving up and shutting it down.
608 # Maximum time a worker can stay alive with no successful
609 # probes before being automatically shut down.
612 # Time after shutting down a worker to retry the
613 # shutdown/destroy operation.
616 # Worker VM image ID.
619 # Tags to add on all resources (VMs, NICs, disks) created by
620 # the container dispatcher. (Arvados's own tags --
621 # InstanceType, IdleBehavior, and InstanceSecret -- will also
626 # Prefix for predefined tags used by Arvados (InstanceSetID,
627 # InstanceType, InstanceSecret, IdleBehavior). With the
628 # default value "Arvados", tags are "ArvadosInstanceSetID",
629 # "ArvadosInstanceSecret", etc.
631 # This should only be changed while no cloud resources are in
632 # use and the cloud dispatcher is not running. Otherwise,
633 # VMs/resources that were added using the old tag prefix will
634 # need to be detected and cleaned up manually.
635 TagKeyPrefix: Arvados
637 # Cloud driver: "azure" (Microsoft Azure) or "ec2" (Amazon AWS).
640 # Cloud-specific driver parameters.
647 # (ec2) Instance configuration.
653 AdminUsername: debian
655 # (azure) Credentials.
661 # (azure) Instance configuration.
662 CloudEnvironment: AzurePublicCloud
669 DeleteDanglingResourcesAfter: 20s
670 AdminUsername: arvados
674 # Use the instance type name as the key (in place of "SAMPLE" in
675 # this sample entry).
677 # Cloud provider's instance type. Defaults to the configured type name.
681 IncludedScratch: 16GB
689 SendUserSetupNotificationEmail: true
691 # Bug/issue report notification to and from addresses
692 IssueReporterEmailFrom: "arvados@example.com"
693 IssueReporterEmailTo: "arvados@example.com"
694 SupportEmailAddress: "arvados@example.com"
696 # Generic issue email from
697 EmailFrom: "arvados@example.com"
706 # API endpoint host or host:port; default is {id}.arvadosapi.com
707 Host: sample.arvadosapi.com
709 # Perform a proxy request when a local client requests an
710 # object belonging to this remote.
713 # Default "https". Can be set to "http" for testing.
716 # Disable TLS verify. Can be set to true for testing.
719 # When users present tokens issued by this remote cluster, and
720 # their accounts are active on the remote cluster, activate
721 # them on this cluster too.
727 ActivationContactLink: mailto:info@arvados.org
728 ArvadosDocsite: https://doc.arvados.org
729 ArvadosPublicDataDocURL: https://playground.arvados.org/projects/public
730 ShowUserAgreementInline: false
733 # Scratch directory used by the remote repository browsing
734 # feature. If it doesn't exist, it (and any missing parents) will be
735 # created using mkdir_p.
736 RepositoryCache: /var/www/arvados-workbench/current/tmp/git
738 # Below is a sample setting of user_profile_form_fields config parameter.
739 # This configuration parameter should be set to either false (to disable) or
740 # to a map as shown below.
741 # Configure the map of input fields to be displayed in the profile page
742 # using the attribute "key" for each of the input fields.
743 # This sample shows configuration with one required and one optional form fields.
744 # For each of these input fields:
745 # You can specify "Type" as "text" or "select".
746 # List the "Options" to be displayed for each of the "select" menu.
747 # Set "Required" as "true" for any of these fields to make them required.
748 # If any of the required fields are missing in the user's profile, the user will be
749 # redirected to the profile page before they can access any Workbench features.
750 UserProfileFormFields: {}
751 # exampleTextValue: # key that will be set in properties
754 # FormFieldDescription: ""
757 # exampleOptionsValue:
760 # FormFieldDescription: ""
768 # Use "UserProfileFormMessage to configure the message you want
769 # to display on the profile page.
770 UserProfileFormMessage: 'Welcome to Arvados. All <span style="color:red">required fields</span> must be completed before you can proceed.'
772 # Mimetypes of applications for which the view icon
773 # would be enabled in a collection's show page.
774 # It is sufficient to list only applications here.
775 # No need to list text and image types.
776 ApplicationMimetypesWithViewIcon:
793 # The maximum number of bytes to load in the log viewer
794 LogViewerMaxBytes: 1M
796 # When anonymous_user_token is configured, show public projects page
797 EnablePublicProjectsPage: true
799 # By default, disable the "Getting Started" popup which is specific to Arvados playground
800 EnableGettingStartedPopup: false
802 # Ask Arvados API server to compress its response payloads.
803 APIResponseCompression: true
805 # Timeouts for API requests.
806 APIClientConnectTimeout: 2m
807 APIClientReceiveTimeout: 5m
809 # Maximum number of historic log records of a running job to fetch
810 # and display in the Log tab, while subscribing to web sockets.
811 RunningJobLogRecordsToFetch: 2000
813 # In systems with many shared projects, loading of dashboard and topnav
814 # cab be slow due to collections indexing; use the following parameters
815 # to suppress these properties
816 ShowRecentCollectionsOnDashboard: true
817 ShowUserNotifications: true
819 # Enable/disable "multi-site search" in top nav ("true"/"false"), or
820 # a link to the multi-site search page on a "home" Workbench site.
823 # https://workbench.qr1hi.arvadosapi.com/collections/multisite
826 # Should workbench allow management of local git repositories? Set to false if
827 # the jobs api is disabled and there are no local git repositories.
830 SiteName: Arvados Workbench
831 ProfilingEnabled: false
833 # This is related to obsolete Google OpenID 1.0 login
834 # but some workbench stuff still expects it to be set.
835 DefaultOpenIdPrefix: "https://www.google.com/accounts/o8/id"
839 FileViewersConfigURL: ""
841 # Use experimental controller code (see https://dev.arvados.org/issues/14287)
842 EnableBetaController14287: false