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"
78 # max concurrent connections per arvados server daemon
81 # All parameters here are passed to the PG client library in a connection string;
82 # see https://www.postgresql.org/docs/current/static/libpq-connect.html#LIBPQ-PARAMKEYWORDS
89 # Maximum size (in bytes) allowed for a single API request. This
90 # limit is published in the discovery document for use by clients.
91 # Note: You must separately configure the upstream web server or
92 # proxy to actually enforce the desired maximum request size on the
94 MaxRequestSize: 134217728
96 # Limit the number of bytes read from the database during an index
97 # request (by retrieving and returning fewer rows than would
98 # normally be returned in a single response).
99 # Note 1: This setting never reduces the number of returned rows to
100 # zero, no matter how big the first data row is.
101 # Note 2: Currently, this is only checked against a specific set of
102 # columns that tend to get large (collections.manifest_text,
103 # containers.mounts, workflows.definition). Other fields (e.g.,
104 # "properties" hashes) are not counted against this limit.
105 MaxIndexDatabaseRead: 134217728
107 # Maximum number of items to return when responding to a APIs that
108 # can return partial result sets using limit and offset parameters
109 # (e.g., *.index, groups.contents). If a request specifies a "limit"
110 # parameter higher than this value, this value is used instead.
111 MaxItemsPerResponse: 1000
113 # API methods to disable. Disabled methods are not listed in the
114 # discovery document, and respond 404 to all requests.
115 # Example: ["jobs.create", "pipeline_instances.create"]
118 # Interval (seconds) between asynchronous permission view updates. Any
119 # permission-updating API called with the 'async' parameter schedules a an
120 # update on the permission view in the future, if not already scheduled.
121 AsyncPermissionsUpdateInterval: 20s
123 # Maximum number of concurrent outgoing requests to make while
124 # serving a single incoming multi-cluster (federated) request.
125 MaxRequestAmplification: 4
127 # RailsSessionSecretToken is a string of alphanumeric characters
128 # used by Rails to sign session tokens. IMPORTANT: This is a
129 # site secret. It should be at least 50 characters.
130 RailsSessionSecretToken: ""
132 # Maximum wall clock time to spend handling an incoming request.
136 # Config parameters to automatically setup new users. If enabled,
137 # this users will be able to self-activate. Enable this if you want
138 # to run an open instance where anyone can create an account and use
139 # the system without requiring manual approval.
141 # The params auto_setup_new_users_with_* are meaningful only when auto_setup_new_users is turned on.
142 # auto_setup_name_blacklist is a list of usernames to be blacklisted for auto setup.
143 AutoSetupNewUsers: false
144 AutoSetupNewUsersWithVmUUID: ""
145 AutoSetupNewUsersWithRepository: false
146 AutoSetupUsernameBlacklist: [arvados, git, gitolite, gitolite-admin, root, syslog]
148 # When new_users_are_active is set to true, new users will be active
149 # immediately. This skips the "self-activate" step which enforces
150 # user agreements. Should only be enabled for development.
151 NewUsersAreActive: false
153 # The e-mail address of the user you would like to become marked as an admin
154 # user on their first login.
155 # In the default configuration, authentication happens through the Arvados SSO
156 # server, which uses OAuth2 against Google's servers, so in that case this
157 # should be an address associated with a Google account.
158 AutoAdminUserWithEmail: ""
160 # If auto_admin_first_user is set to true, the first user to log in when no
161 # other admin users exist will automatically become an admin user.
162 AutoAdminFirstUser: false
164 # Email address to notify whenever a user creates a profile for the
166 UserProfileNotificationAddress: ""
167 AdminNotifierEmailFrom: arvados@example.com
168 EmailSubjectPrefix: "[ARVADOS] "
169 UserNotifierEmailFrom: arvados@example.com
170 NewUserNotificationRecipients: []
171 NewInactiveUserNotificationRecipients: []
174 # Time to keep audit logs, in seconds. (An audit log is a row added
175 # to the "logs" table in the PostgreSQL database each time an
176 # Arvados object is created, modified, or deleted.)
178 # Currently, websocket event notifications rely on audit logs, so
179 # this should not be set lower than 300 (5 minutes).
182 # Maximum number of log rows to delete in a single SQL transaction.
184 # If max_audit_log_delete_batch is 0, log entries will never be
185 # deleted by Arvados. Cleanup can be done by an external process
186 # without affecting any Arvados system processes, as long as very
187 # recent (<5 minutes old) logs are not deleted.
189 # 100000 is a reasonable batch size for most sites.
192 # Attributes to suppress in events and audit logs. Notably,
193 # specifying ["manifest_text"] here typically makes the database
194 # smaller and faster.
196 # Warning: Using any non-empty value here can have undesirable side
197 # effects for any client or component that relies on event logs.
198 # Use at your own risk.
199 UnloggedAttributes: []
203 # Logging threshold: panic, fatal, error, warn, info, debug, or
207 # Logging format: json or text
210 # Maximum characters of (JSON-encoded) query parameters to include
211 # in each request log entry. When params exceed this size, they will
212 # be JSON-encoded, truncated to this size, and logged as
214 MaxRequestLogParamsSize: 2000
217 # Allow clients to create collections by providing a manifest with
218 # unsigned data blob locators. IMPORTANT: This effectively disables
219 # access controls for data stored in Keep: a client who knows a hash
220 # can write a manifest that references the hash, pass it to
221 # collections.create (which will create a permission link), use
222 # collections.get to obtain a signature for that data locator, and
223 # use that signed locator to retrieve the data from Keep. Therefore,
224 # do not turn this on if your users expect to keep data private from
228 # blob_signing_key is a string of alphanumeric characters used to
229 # generate permission signatures for Keep locators. It must be
230 # identical to the permission key given to Keep. IMPORTANT: This is
231 # a site secret. It should be at least 50 characters.
233 # Modifying blob_signing_key will invalidate all existing
234 # signatures, which can cause programs to fail (e.g., arv-put,
235 # arv-get, and Crunch jobs). To avoid errors, rotate keys only when
236 # no such processes are running.
239 # Default replication level for collections. This is used when a
240 # collection's replication_desired attribute is nil.
241 DefaultReplication: 2
243 # Lifetime (in seconds) of blob permission signatures generated by
244 # the API server. This determines how long a client can take (after
245 # retrieving a collection record) to retrieve the collection data
246 # from Keep. If the client needs more time than that (assuming the
247 # collection still has the same content and the relevant user/token
248 # still has permission) the client can retrieve the collection again
249 # to get fresh signatures.
251 # This must be exactly equal to the -blob-signature-ttl flag used by
252 # keepstore servers. Otherwise, reading data blocks and saving
253 # collections will fail with HTTP 403 permission errors.
255 # Modifying blob_signature_ttl invalidates existing signatures; see
256 # blob_signing_key note above.
258 # The default is 2 weeks.
261 # Default lifetime for ephemeral collections: 2 weeks. This must not
262 # be less than blob_signature_ttl.
263 DefaultTrashLifetime: 336h
265 # Interval (seconds) between trash sweeps. During a trash sweep,
266 # collections are marked as trash if their trash_at time has
267 # arrived, and deleted if their delete_at time has arrived.
268 TrashSweepInterval: 60s
270 # If true, enable collection versioning.
271 # When a collection's preserve_version field is true or the current version
272 # is older than the amount of seconds defined on preserve_version_if_idle,
273 # a snapshot of the collection's previous state is created and linked to
274 # the current collection.
275 CollectionVersioning: false
277 # 0s = auto-create a new version on every update.
278 # -1s = never auto-create new versions.
279 # > 0s = auto-create a new version when older than the specified number of seconds.
280 PreserveVersionIfIdle: -1s
283 # These settings are provided by your OAuth2 provider (e.g.,
285 ProviderAppSecret: ""
289 # Git repositories must be readable by api server, or you won't be
290 # able to submit crunch jobs. To pass the test suites, put a clone
291 # of the arvados tree in {git_repositories_dir}/arvados.git or
292 # {git_repositories_dir}/arvados/.git
293 Repositories: /var/lib/arvados/git/repositories
301 # List of supported Docker Registry image formats that compute nodes
302 # are able to use. `arv keep docker` will error out if a user tries
303 # to store an image with an unsupported format. Use an empty array
304 # to skip the compatibility check (and display a warning message to
307 # Example for sites running docker < 1.10: ["v1"]
308 # Example for sites running docker >= 1.10: ["v2"]
309 # Example for disabling check: []
310 SupportedDockerImageFormats: ["v2"]
312 # Include details about job reuse decisions in the server log. This
313 # causes additional database queries to run, so it should not be
314 # enabled unless you expect to examine the resulting logs for
315 # troubleshooting purposes.
316 LogReuseDecisions: false
318 # Default value for keep_cache_ram of a container's runtime_constraints.
319 DefaultKeepCacheRAM: 268435456
321 # Number of times a container can be unlocked before being
322 # automatically cancelled.
323 MaxDispatchAttempts: 5
325 # Default value for container_count_max for container requests. This is the
326 # number of times Arvados will create a new container to satisfy a container
327 # request. If a container is cancelled it will retry a new container if
328 # container_count < container_count_max on any container requests associated
329 # with the cancelled container.
332 # The maximum number of compute nodes that can be in use simultaneously
333 # If this limit is reduced, any existing nodes with slot number >= new limit
334 # will not be counted against the new limit. In other words, the new limit
335 # won't be strictly enforced until those nodes with higher slot numbers
339 # Preemptible instance support (e.g. AWS Spot Instances)
340 # When true, child containers will get created with the preemptible
341 # scheduling parameter parameter set.
342 UsePreemptibleInstances: false
344 # PEM encoded SSH key (RSA, DSA, or ECDSA) used by the
345 # (experimental) cloud dispatcher for executing containers on
346 # worker VMs. Begins with "-----BEGIN RSA PRIVATE KEY-----\n"
347 # and ends with "\n-----END RSA PRIVATE KEY-----\n".
348 DispatchPrivateKey: none
350 # Maximum time to wait for workers to come up before abandoning
351 # stale locks from a previous dispatch process.
355 # When you run the db:delete_old_container_logs task, it will find
356 # containers that have been finished for at least this many seconds,
357 # and delete their stdout, stderr, arv-mount, crunch-run, and
358 # crunchstat logs from the logs table.
361 # These two settings control how frequently log events are flushed to the
362 # database. Log lines are buffered until either crunch_log_bytes_per_event
363 # has been reached or crunch_log_seconds_between_events has elapsed since
365 LogBytesPerEvent: 4096
366 LogSecondsBetweenEvents: 1
368 # The sample period for throttling logs.
369 LogThrottlePeriod: 60s
371 # Maximum number of bytes that job can log over crunch_log_throttle_period
372 # before being silenced until the end of the period.
373 LogThrottleBytes: 65536
375 # Maximum number of lines that job can log over crunch_log_throttle_period
376 # before being silenced until the end of the period.
377 LogThrottleLines: 1024
379 # Maximum bytes that may be logged by a single job. Log bytes that are
380 # silenced by throttling are not counted against this total.
381 LimitLogBytesPerJob: 67108864
383 LogPartialLineThrottlePeriod: 5s
385 # Container logs are written to Keep and saved in a
386 # collection, which is updated periodically while the
387 # container runs. This value sets the interval between
388 # collection updates.
391 # The log collection is also updated when the specified amount of
392 # log data (given in bytes) is produced in less than one update
398 # Path to dns server configuration directory
399 # (e.g. /etc/unbound.d/conf.d). If false, do not write any config
400 # files or touch restart.txt (see below).
403 # Template file for the dns server host snippets. See
404 # unbound.template in this directory for an example. If false, do
405 # not write any config files.
406 DNSServerConfTemplate: ""
408 # String to write to {dns_server_conf_dir}/restart.txt (with a
409 # trailing newline) after updating local data. If false, do not
410 # open or write the restart.txt file.
411 DNSServerReloadCommand: ""
413 # Command to run after each DNS update. Template variables will be
414 # substituted; see the "unbound" example below. If false, do not run
416 DNSServerUpdateCommand: ""
418 ComputeNodeDomain: ""
419 ComputeNodeNameservers:
422 # Hostname to assign to a compute node when it sends a "ping" and the
423 # hostname in its Node record is nil.
424 # During bootstrapping, the "ping" script is expected to notice the
425 # hostname given in the ping response, and update its unix hostname
427 # If false, leave the hostname alone (this is appropriate if your compute
428 # nodes' hostnames are already assigned by some other mechanism).
430 # One way or another, the hostnames of your node records should agree
431 # with your DNS records and your /etc/slurm-llnl/slurm.conf files.
433 # Example for compute0000, compute0001, ....:
434 # assign_node_hostname: compute%<slot_number>04d
435 # (See http://ruby-doc.org/core-2.2.2/Kernel.html#method-i-format for more.)
436 AssignNodeHostname: "compute%<slot_number>d"
439 # Enable the legacy Jobs API. This value must be a string.
440 # 'auto' -- (default) enable the Jobs API only if it has been used before
441 # (i.e., there are job records in the database)
442 # 'true' -- enable the Jobs API despite lack of existing records.
443 # 'false' -- disable the Jobs API despite presence of existing records.
446 # Git repositories must be readable by api server, or you won't be
447 # able to submit crunch jobs. To pass the test suites, put a clone
448 # of the arvados tree in {git_repositories_dir}/arvados.git or
449 # {git_repositories_dir}/arvados/.git
450 GitInternalDir: /var/lib/arvados/internal.git
452 # Docker image to be used when none found in runtime_constraints of a job
453 DefaultDockerImage: ""
455 # none or slurm_immediate
456 CrunchJobWrapper: none
458 # username, or false = do not set uid when running jobs.
459 CrunchJobUser: crunch
461 # The web service must be able to create/write this file, and
462 # crunch-job must be able to stat() it.
463 CrunchRefreshTrigger: /tmp/crunch_refresh_trigger
465 # Control job reuse behavior when two completed jobs match the
466 # search criteria and have different outputs.
468 # If true, in case of a conflict, reuse the earliest job (this is
469 # similar to container reuse behavior).
471 # If false, in case of a conflict, do not reuse any completed job,
472 # but do reuse an already-running job if available (this is the
473 # original job reuse behavior, and is still the default).
474 ReuseJobIfOutputsDiffer: false
477 # Enable the cloud scheduler (experimental).
480 # Name/number of port where workers' SSH services listen.
483 # Interval between queue polls.
486 # Shell command to execute on each worker to determine whether
487 # the worker is booted and ready to run containers. It should
488 # exit zero if the worker is ready.
489 BootProbeCommand: "docker ps"
491 # Minimum interval between consecutive probes to a single
495 # Maximum probes per second, across all workers in a pool.
496 MaxProbesPerSecond: 10
498 # Time before repeating SIGTERM when killing a container.
501 # Time to give up on SIGTERM and write off the worker.
504 # Maximum create/destroy-instance operations per second (0 =
506 MaxCloudOpsPerSecond: 0
508 # Interval between cloud provider syncs/updates ("list all
512 # Time to leave an idle worker running (in case new containers
513 # appear in the queue that it can run) before shutting it
517 # Time to wait for a new worker to boot (i.e., pass
518 # BootProbeCommand) before giving up and shutting it down.
521 # Maximum time a worker can stay alive with no successful
522 # probes before being automatically shut down.
525 # Time after shutting down a worker to retry the
526 # shutdown/destroy operation.
529 # Worker VM image ID.
532 # Tags to add on all resources (VMs, NICs, disks) created by
533 # the container dispatcher. (Arvados's own tags --
534 # InstanceType, IdleBehavior, and InstanceSecret -- will also
539 # Prefix for predefined tags used by Arvados (InstanceSetID,
540 # InstanceType, InstanceSecret, IdleBehavior). With the
541 # default value "Arvados", tags are "ArvadosInstanceSetID",
542 # "ArvadosInstanceSecret", etc.
544 # This should only be changed while no cloud resources are in
545 # use and the cloud dispatcher is not running. Otherwise,
546 # VMs/resources that were added using the old tag prefix will
547 # need to be detected and cleaned up manually.
548 TagKeyPrefix: Arvados
550 # Cloud driver: "azure" (Microsoft Azure) or "ec2" (Amazon AWS).
553 # Cloud-specific driver parameters.
560 # (ec2) Instance configuration.
566 AdminUsername: debian
568 # (azure) Credentials.
574 # (azure) Instance configuration.
575 CloudEnvironment: AzurePublicCloud
582 DeleteDanglingResourcesAfter: 20s
583 AdminUsername: arvados
587 # Use the instance type name as the key (in place of "SAMPLE" in
588 # this sample entry).
590 # Cloud provider's instance type. Defaults to the configured type name.
594 IncludedScratch: 16GB
602 SendUserSetupNotificationEmail: ""
603 IssueReporterEmailFrom: ""
604 IssueReporterEmailTo: ""
605 SupportEmailAddress: ""
615 # API endpoint host or host:port; default is {id}.arvadosapi.com
616 Host: sample.arvadosapi.com
618 # Perform a proxy request when a local client requests an
619 # object belonging to this remote.
622 # Default "https". Can be set to "http" for testing.
625 # Disable TLS verify. Can be set to true for testing.
628 # When users present tokens issued by this remote cluster, and
629 # their accounts are active on the remote cluster, activate
630 # them on this cluster too.
636 ActivationContactLink: mailto:info@arvados.org
637 ArvadosDocsite: https://doc.arvados.org
638 ArvadosPublicDataDocURL: https://playground.arvados.org/projects/public
639 ShowUserAgreementInline: false
642 RepositoryCache: /var/www/arvados-workbench/current/tmp/git
643 UserProfileFormFields:
647 FormFieldDescription: ""
649 UserProfileFormMessage: 'Welcome to Arvados. All <span style="color:red">required fields</span> must be completed before you can proceed.'
650 ApplicationMimetypesWithViewIcon:
666 LogViewerMaxBytes: 1M
667 EnablePublicProjectsPage: true
668 EnableGettingStartedPopup: false
669 APIResponseCompression: true
670 APIClientConnectTimeout: 2m
671 APIClientReceiveTimeout: 5m
672 RunningJobLogRecordsToFetch: 2000
673 ShowRecentCollectionsOnDashboard: true
674 ShowUserNotifications: true
675 MultiSiteSearch: false
677 SiteName: Arvados Workbench
681 FileViewersConfigURL: ""
683 # Use experimental controller code (see https://dev.arvados.org/issues/14287)
684 EnableBetaController14287: false