The Arvados configuration is stored at /etc/arvados/config.yml
# Copyright (C) The Arvados Authors. All rights reserved. # # SPDX-License-Identifier: AGPL-3.0 # Do not use this file for site configuration. Create # /etc/arvados/config.yml instead. # # The order of precedence (highest to lowest): # 1. Legacy component-specific config files (deprecated) # 2. /etc/arvados/config.yml # 3. config.default.yml Clusters: xxxxx: # Token used internally by Arvados components to authenticate to # one another. Use a string of at least 50 random alphanumerics. SystemRootToken: "" # Token to be included in all healthcheck requests. Disabled by default. # Server expects request header of the format "Authorization: Bearer xxx" ManagementToken: "" Services: # Each of the service sections below specifies InternalURLs # (each with optional ListenURL) and ExternalURL. # # InternalURLs specify how other Arvados service processes will # connect to the service. Typically these use internal hostnames # and high port numbers. Example: # # InternalURLs: # "http://host1.internal.example:12345": {} # "http://host2.internal.example:12345": {} # # ListenURL specifies the address and port the service process's # HTTP server should listen on, if different from the # InternalURL itself. Example, using an intermediate TLS proxy: # # InternalURLs: # "https://host1.internal.example": # ListenURL: "http://10.0.0.7:12345" # # When there are multiple InternalURLs configured, the service # process will try listening on each InternalURLs (using # ListenURL if provided) until one works. If you use a ListenURL # like "0.0.0.0" which can be bound on any machine, use an # environment variable # ARVADOS_SERVICE_INTERNAL_URL=http://host1.internal.example to # control which entry to use. # # ExternalURL specifies how applications/clients will connect to # the service, regardless of whether they are inside or outside # the cluster. Example: # # ExternalURL: "https://keep.zzzzz.example.com/" # # To avoid routing internal traffic through external networks, # use split-horizon DNS for ExternalURL host names: inside the # cluster's private network "host.zzzzz.example.com" resolves to # the host's private IP address, while outside the cluster # "host.zzzzz.example.com" resolves to the host's public IP # address (or its external gateway or load balancer). RailsAPI: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" Controller: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" Websocket: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" Keepbalance: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" DispatchCloud: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" DispatchLSF: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" DispatchSLURM: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" Keepproxy: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" WebDAV: InternalURLs: {SAMPLE: {ListenURL: ""}} # Base URL for Workbench inline preview. If blank, use # WebDAVDownload instead, and disable inline preview. # If both are empty, downloading collections from workbench # will be impossible. # # It is important to properly configure the download service # to migitate cross-site-scripting (XSS) attacks. A HTML page # can be stored in collection. If an attacker causes a victim # to visit that page through Workbench, it will be rendered by # the browser. If all collections are served at the same # domain, the browser will consider collections as coming from # the same origin and having access to the same browsing data, # enabling malicious Javascript on that page to access Arvados # on behalf of the victim. # # This is mitigating by having separate domains for each # collection, or limiting preview to circumstances where the # collection is not accessed with the user's regular # full-access token. # # Serve preview links using uuid or pdh in subdomain # (requires wildcard DNS and TLS certificate) # https://*.collections.uuid_prefix.arvadosapi.com # # Serve preview links using uuid or pdh in main domain # (requires wildcard DNS and TLS certificate) # https://*--collections.uuid_prefix.arvadosapi.com # # Serve preview links by setting uuid or pdh in the path. # This configuration only allows previews of public data or # collection-sharing links, because these use the anonymous # user token or the token is already embedded in the URL. # Other data must be handled as downloads via WebDAVDownload: # https://collections.uuid_prefix.arvadosapi.com # ExternalURL: "" WebDAVDownload: InternalURLs: {SAMPLE: {ListenURL: ""}} # Base URL for download links. If blank, serve links to WebDAV # with disposition=attachment query param. Unlike preview links, # browsers do not render attachments, so there is no risk of XSS. # # If WebDAVDownload is blank, and WebDAV uses a # single-origin form, then Workbench will show an error page # # Serve download links by setting uuid or pdh in the path: # https://download.uuid_prefix.arvadosapi.com # ExternalURL: "" Keepstore: InternalURLs: SAMPLE: ListenURL: "" # Rendezvous is normally empty/omitted. When changing the # URL of a Keepstore service, Rendezvous should be set to # the old URL (with trailing slash omitted) to preserve # rendezvous ordering. Rendezvous: "" ExternalURL: "" Composer: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" WebShell: InternalURLs: {SAMPLE: {ListenURL: ""}} # ShellInABox service endpoint URL for a given VM. If empty, do not # offer web shell logins. # # E.g., using a path-based proxy server to forward connections to shell hosts: # https://webshell.uuid_prefix.arvadosapi.com # # E.g., using a name-based proxy server to forward connections to shell hosts: # https://*.webshell.uuid_prefix.arvadosapi.com ExternalURL: "" Workbench1: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" Workbench2: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" Health: InternalURLs: {SAMPLE: {ListenURL: ""}} ExternalURL: "" PostgreSQL: # max concurrent connections per arvados server daemon ConnectionPool: 32 Connection: # All parameters here are passed to the PG client library in a connection string; # see https://www.postgresql.org/docs/current/static/libpq-connect.html#LIBPQ-PARAMKEYWORDS host: "" port: "" user: "" password: "" dbname: "" SAMPLE: "" API: # Limits for how long a client token created by regular users can be valid, # and also is used as a default expiration policy when no expiration date is # specified. # Default value zero means token expirations don't get clamped and no # default expiration is set. MaxTokenLifetime: 0s # Maximum size (in bytes) allowed for a single API request. This # limit is published in the discovery document for use by clients. # Note: You must separately configure the upstream web server or # proxy to actually enforce the desired maximum request size on the # server side. MaxRequestSize: 134217728 # Limit the number of bytes read from the database during an index # request (by retrieving and returning fewer rows than would # normally be returned in a single response). # Note 1: This setting never reduces the number of returned rows to # zero, no matter how big the first data row is. # Note 2: Currently, this is only checked against a specific set of # columns that tend to get large (collections.manifest_text, # containers.mounts, workflows.definition). Other fields (e.g., # "properties" hashes) are not counted against this limit. MaxIndexDatabaseRead: 134217728 # Maximum number of items to return when responding to a APIs that # can return partial result sets using limit and offset parameters # (e.g., *.index, groups.contents). If a request specifies a "limit" # parameter higher than this value, this value is used instead. MaxItemsPerResponse: 1000 # Maximum number of concurrent requests to process concurrently # in a single service process, or 0 for no limit. # # Note this applies to all Arvados services (controller, webdav, # websockets, etc.). Concurrency in the controller service is # also effectively limited by MaxConcurrentRailsRequests (see # below) because most controller requests proxy through to the # RailsAPI service. # # HTTP proxies and load balancers downstream of arvados services # should be configured to allow at least {MaxConcurrentRequest + # MaxQueuedRequests + MaxGatewayTunnels} concurrent requests. MaxConcurrentRequests: 64 # Maximum number of concurrent requests to process concurrently # in a single RailsAPI service process, or 0 for no limit. MaxConcurrentRailsRequests: 8 # Maximum number of incoming requests to hold in a priority # queue waiting for one of the MaxConcurrentRequests slots to be # free. When the queue is longer than this, respond 503 to the # lowest priority request. # # If MaxQueuedRequests is 0, respond 503 immediately to # additional requests while at the MaxConcurrentRequests limit. MaxQueuedRequests: 128 # Maximum time a "lock container" request is allowed to wait in # the incoming request queue before returning 503. MaxQueueTimeForLockRequests: 2s # Maximum number of active gateway tunnel connections. One slot # is consumed by each "container shell" connection. If using an # HPC dispatcher (LSF or Slurm), one slot is consumed by each # running container. These do not count toward # MaxConcurrentRequests. MaxGatewayTunnels: 1000 # Maximum number of 64MiB memory buffers per Keepstore server process, or # 0 for no limit. When this limit is reached, up to # (MaxConcurrentRequests - MaxKeepBlobBuffers) HTTP requests requiring # buffers (like GET and PUT) will wait for buffer space to be released. # Any HTTP requests beyond MaxConcurrentRequests will receive an # immediate 503 response. # # MaxKeepBlobBuffers should be set such that (MaxKeepBlobBuffers * 64MiB # * 1.1) fits comfortably in memory. On a host dedicated to running # Keepstore, divide total memory by 88MiB to suggest a suitable value. # For example, if grep MemTotal /proc/meminfo reports MemTotal: 7125440 # kB, compute 7125440 / (88 * 1024)=79 and set MaxKeepBlobBuffers: 79 MaxKeepBlobBuffers: 128 # API methods to disable. Disabled methods are not listed in the # discovery document, and respond 404 to all requests. # Example: {"jobs.create":{}, "pipeline_instances.create": {}} DisabledAPIs: {} # Interval (seconds) between asynchronous permission view updates. Any # permission-updating API called with the 'async' parameter schedules a an # update on the permission view in the future, if not already scheduled. AsyncPermissionsUpdateInterval: 20s # Maximum number of concurrent outgoing requests to make while # serving a single incoming multi-cluster (federated) request. MaxRequestAmplification: 4 # Maximum wall clock time to spend handling an incoming request. RequestTimeout: 5m # Websocket will send a periodic empty event after 'SendTimeout' # if there is no other activity to maintain the connection / # detect dropped connections. SendTimeout: 60s WebsocketClientEventQueue: 64 WebsocketServerEventQueue: 4 # Timeout on requests to internal Keep services. KeepServiceRequestTimeout: 15s # Vocabulary file path, local to the node running the controller. # This JSON file should contain the description of what's allowed # as object's metadata. Its format is described at: # https://doc.arvados.org/admin/metadata-vocabulary.html VocabularyPath: "" # If true, a project must have a non-empty description field in # order to be frozen. FreezeProjectRequiresDescription: false # Project properties that must have non-empty values in order to # freeze a project. Example: "property_name": {} FreezeProjectRequiresProperties: SAMPLE: {} # If true, only an admin user can un-freeze a project. If false, # any user with "manage" permission can un-freeze. UnfreezeProjectRequiresAdmin: false # (Experimental) Use row-level locking on update API calls. LockBeforeUpdate: false Users: # Config parameters to automatically setup new users. If enabled, # this users will be able to self-activate. Enable this if you want # to run an open instance where anyone can create an account and use # the system without requiring manual approval. # # The params AutoSetupNewUsersWith* are meaningful only when AutoSetupNewUsers is turned on. # AutoSetupUsernameBlacklist is a list of usernames to be blacklisted for auto setup. AutoSetupNewUsers: false AutoSetupNewUsersWithVmUUID: "" AutoSetupUsernameBlacklist: arvados: {} git: {} gitolite: {} gitolite-admin: {} root: {} syslog: {} SAMPLE: {} # When NewUsersAreActive is set to true, new users will be active # immediately. This skips the "self-activate" step which enforces # user agreements. Should only be enabled for development. NewUsersAreActive: false # Newly activated users (whether set up by an admin or via # AutoSetupNewUsers) immediately become visible to other active # users. # # On a multi-tenant cluster, where the intent is for users to be # invisible to one another unless they have been added to the # same group(s) via Workbench admin interface, change this to # false. ActivatedUsersAreVisibleToOthers: true # If a user creates an account with this email address, they # will be automatically set to admin. AutoAdminUserWithEmail: "" # If AutoAdminFirstUser is set to true, the first user to log in when no # other admin users exist will automatically become an admin user. AutoAdminFirstUser: false # Support email address to display in Workbench. SupportEmailAddress: "arvados@example.com" # Outgoing email configuration: # # In order to send mail, Arvados expects a default SMTP server # on localhost:25. It cannot require authentication on # connections from localhost. That server should be configured # to relay mail to a "real" SMTP server that is able to send # email on behalf of your domain. # Recipient for notification email sent out when a user sets a # profile on their account. UserProfileNotificationAddress: "" # When sending a NewUser, NewInactiveUser, or UserProfile # notification, this is the 'From' address to use AdminNotifierEmailFrom: arvados@example.com # Prefix for email subjects for NewUser and NewInactiveUser emails EmailSubjectPrefix: "[ARVADOS] " # When sending a welcome email to the user, the 'From' address to use UserNotifierEmailFrom: arvados@example.com # The welcome email sent to new users will be blind copied to # these addresses. UserNotifierEmailBcc: SAMPLE: {} # Recipients for notification email sent out when a user account # is created and already set up to be able to log in NewUserNotificationRecipients: SAMPLE: {} # Recipients for notification email sent out when a user account # has been created but the user cannot log in until they are # set up by an admin. NewInactiveUserNotificationRecipients: SAMPLE: {} # Set AnonymousUserToken to enable anonymous user access. Populate this # field with a random string at least 50 characters long. AnonymousUserToken: "" # The login provider for a user may supply a primary email # address and one or more alternate email addresses. If a new # user has an alternate email address with the domain given # here, use the username from the alternate email to generate # the user's Arvados username. Otherwise, the username from # user's primary email address is used for the Arvados username. # Currently implemented for OpenID Connect only. PreferDomainForUsername: "" # Send an email to each user when their account has been set up # (meaning they are able to log in). SendUserSetupNotificationEmail: true # Ruby ERB template used for the email sent out to users when # they have been set up. UserSetupMailText: | <% if not @user.full_name.empty? -%> <%= @user.full_name %>, <% else -%> Hi there, <% end -%> Your Arvados account has been set up. You can log in at <%= Rails.configuration.Services.Workbench1.ExternalURL %> Thanks, Your Arvados administrator. # If RoleGroupsVisibleToAll is true, all role groups are visible # to all active users. # # If false, users must be granted permission to role groups in # order to see them. This is more appropriate for a multi-tenant # cluster. RoleGroupsVisibleToAll: true # If CanCreateRoleGroups is true, regular (non-admin) users can # create new role groups. # # If false, only admins can create new role groups. CanCreateRoleGroups: true # During each period, a log entry with event_type="activity" # will be recorded for each user who is active during that # period. The object_uuid attribute will indicate the user's # UUID. # # Multiple log entries for the same user may be generated during # a period if there are multiple controller processes or a # controller process is restarted. # # Use 0 to disable activity logging. ActivityLoggingPeriod: 24h # The SyncUser* options control what system resources are managed by # arvados-login-sync on shell nodes. They correspond to: # * SyncUserAccounts: The user's Unix account on the shell node # * SyncUserGroups: The group memberships of that account # * SyncUserSSHKeys: Whether to authorize the user's Arvados SSH keys # * SyncUserAPITokens: Whether to set up the user's Arvados API token # All default to true. SyncUserAccounts: true SyncUserGroups: true SyncUserSSHKeys: true SyncUserAPITokens: true # If SyncUserGroups=true, then arvados-login-sync will ensure that all # managed accounts are members of the Unix groups listed in # SyncRequiredGroups, in addition to any groups listed in their Arvados # login permission. The default list includes the "fuse" group so # users can use arv-mount. You can require no groups by specifying an # empty list (i.e., `SyncRequiredGroups: []`). SyncRequiredGroups: - fuse # SyncIgnoredGroups is a list of group names. arvados-login-sync will # never modify these groups. If user login permissions list any groups # in SyncIgnoredGroups, they will be ignored. If a user's Unix account # belongs to any of these groups, arvados-login-sync will not remove # the account from that group. The default is a set of particularly # security-sensitive groups across Debian- and Red Hat-based # distributions. SyncIgnoredGroups: - adm - disk - kmem - mem - root - shadow - staff - sudo - sys - utempter - utmp - wheel AuditLogs: # Time to keep audit logs, in seconds. (An audit log is a row added # to the "logs" table in the PostgreSQL database each time an # Arvados object is created, modified, or deleted.) # # Currently, websocket event notifications rely on audit logs, so # this should not be set lower than 300 (5 minutes). MaxAge: 336h # Maximum number of log rows to delete in a single SQL transaction. # # If MaxDeleteBatch is 0, log entries will never be # deleted by Arvados. Cleanup can be done by an external process # without affecting any Arvados system processes, as long as very # recent (<5 minutes old) logs are not deleted. # # 100000 is a reasonable batch size for most sites. MaxDeleteBatch: 0 # Attributes to suppress in events and audit logs. Notably, # specifying {"manifest_text": {}} here typically makes the database # smaller and faster. # # Warning: Using any non-empty value here can have undesirable side # effects for any client or component that relies on event logs. # Use at your own risk. UnloggedAttributes: {} SystemLogs: # Logging threshold: panic, fatal, error, warn, info, debug, or # trace LogLevel: info # Logging format: json or text Format: json # Maximum characters of (JSON-encoded) query parameters to include # in each request log entry. When params exceed this size, they will # be JSON-encoded, truncated to this size, and logged as # params_truncated. MaxRequestLogParamsSize: 2000 # In all services except RailsAPI, periodically check whether # the incoming HTTP request queue is nearly full (see # MaxConcurrentRequests) and, if so, write a snapshot of the # request queue to {service}-requests.json in the specified # directory. # # Leave blank to disable. RequestQueueDumpDirectory: "" Collections: # Enable access controls for data stored in Keep. This should # always be set to true on a production cluster. BlobSigning: true # BlobSigningKey is a string of alphanumeric characters used to # generate permission signatures for Keep locators. It must be # identical to the permission key given to Keep. IMPORTANT: This # is a site secret. It should be at least 50 characters. # # Modifying BlobSigningKey will invalidate all existing # signatures, which can cause programs to fail (e.g., arv-put, # arv-get, and Crunch jobs). To avoid errors, rotate keys only # when no such processes are running. BlobSigningKey: "" # Enable garbage collection of unreferenced blobs in Keep. BlobTrash: true # Time to leave unreferenced blobs in "trashed" state before # deleting them, or 0 to skip the "trashed" state entirely and # delete unreferenced blobs. # # If you use any Amazon S3 buckets as storage volumes, this # must be at least 24h to avoid occasional data loss. BlobTrashLifetime: 336h # How often to check for (and delete) trashed blocks whose # BlobTrashLifetime has expired. BlobTrashCheckInterval: 24h # Maximum number of concurrent "trash blob" and "delete trashed # blob" operations conducted by a single keepstore process. Each # of these can be set to 0 to disable the respective operation. # # If BlobTrashLifetime is zero, "trash" and "delete trash" # happen at once, so only the lower of these two values is used. BlobTrashConcurrency: 4 BlobDeleteConcurrency: 4 # Maximum number of concurrent "create additional replica of # existing blob" operations conducted by a single keepstore # process. BlobReplicateConcurrency: 4 # Default replication level for collections. This is used when a # collection's replication_desired attribute is nil. DefaultReplication: 2 # BlobSigningTTL determines the minimum lifetime of transient # data, i.e., blocks that are not referenced by # collections. Unreferenced blocks exist for two reasons: # # 1) A data block must be written to a disk/cloud backend device # before a collection can be created/updated with a reference to # it. # # 2) Deleting or updating a collection can remove the last # remaining reference to a data block. # # If BlobSigningTTL is too short, long-running # processes/containers will fail when they take too long (a) # between writing blocks and writing collections that reference # them, or (b) between reading collections and reading the # referenced blocks. # # If BlobSigningTTL is too long, data will still be stored long # after the referring collections are deleted, and you will # needlessly fill up disks or waste money on cloud storage. # # Modifying BlobSigningTTL invalidates existing signatures; see # BlobSigningKey note above. # # The default is 2 weeks. BlobSigningTTL: 336h # When running keep-balance, this is the destination filename for # the list of lost block hashes if there are any, one per line. # Updated automically during each successful run. BlobMissingReport: "" # keep-balance operates periodically, i.e.: do a # scan/balance operation, sleep, repeat. # # BalancePeriod determines the interval between start times of # successive scan/balance operations. If a scan/balance operation # takes longer than BalancePeriod, the next one will follow it # immediately. # # If SIGUSR1 is received during an idle period between operations, # the next operation will start immediately. BalancePeriod: 6h # Limits the number of collections retrieved by keep-balance per # API transaction. If this is zero, page size is # determined by the API server's own page size limits (see # API.MaxItemsPerResponse and API.MaxIndexDatabaseRead). BalanceCollectionBatch: 0 # The size of keep-balance's internal queue of # collections. Higher values may improve throughput by allowing # keep-balance to fetch collections from the database while the # current collection are still being processed, at the expense of # using more memory. If this is zero or omitted, pages are # processed serially. BalanceCollectionBuffers: 4 # Maximum time for a rebalancing run. This ensures keep-balance # eventually gives up and retries if, for example, a network # error causes a hung connection that is never closed by the # OS. It should be long enough that it doesn't interrupt a # long-running balancing operation. BalanceTimeout: 6h # Maximum number of replication_confirmed / # storage_classes_confirmed updates to write to the database # after a rebalancing run. When many updates are needed, this # spreads them over a few runs rather than applying them all at # once. BalanceUpdateLimit: 100000 # Maximum number of "pull block from other server" and "trash # block" requests to send to each keepstore server at a # time. Smaller values use less memory in keepstore and # keep-balance. Larger values allow more progress per # keep-balance iteration. A zero value computes all of the # needed changes but does not apply any. BalancePullLimit: 100000 BalanceTrashLimit: 100000 # Default lifetime for ephemeral collections: 2 weeks. This must not # be less than BlobSigningTTL. DefaultTrashLifetime: 336h # Interval (seconds) between trash sweeps. During a trash sweep, # collections are marked as trash if their trash_at time has # arrived, and deleted if their delete_at time has arrived. TrashSweepInterval: 60s # If true, enable collection versioning. # When a collection's preserve_version field is true or the current version # is older than the amount of seconds defined on PreserveVersionIfIdle, # a snapshot of the collection's previous state is created and linked to # the current collection. CollectionVersioning: true # 0s = auto-create a new version on every update. # -1s = never auto-create new versions. # > 0s = auto-create a new version when older than the specified number of seconds. PreserveVersionIfIdle: 10s # If non-empty, allow project and collection names to contain # the "/" character (slash/stroke/solidus), and replace "/" with # the given string in the filesystem hierarchy presented by # WebDAV. Example values are "%2f" and "{slash}". Names that # contain the substitution string itself may result in confusing # behavior, so a value like "_" is not recommended. # # If the default empty value is used, the server will reject # requests to create or rename a collection when the new name # contains "/". # # If the value "/" is used, project and collection names # containing "/" will be allowed, but they will not be # accessible via WebDAV. # # Use of this feature is not recommended, if it can be avoided. ForwardSlashNameSubstitution: "" # Include "folder objects" in S3 ListObjects responses. S3FolderObjects: true # Managed collection properties. At creation time, if the client didn't # provide the listed keys, they will be automatically populated following # one of the following behaviors: # # * UUID of the user who owns the containing project. # responsible_person_uuid: {Function: original_owner, Protected: true} # # * Default concrete value. # foo_bar: {Value: baz, Protected: false} # # If Protected is true, only an admin user can modify its value. ManagedProperties: SAMPLE: {Function: original_owner, Protected: true} # In "trust all content" mode, Workbench will redirect download # requests to WebDAV preview link, even in the cases when # WebDAV would have to expose XSS vulnerabilities in order to # handle the redirect (see discussion on Services.WebDAV). # # This setting has no effect in the recommended configuration, where the # WebDAV service is configured to have a separate domain for every # collection and XSS protection is provided by browsers' same-origin # policy. # # The default setting (false) is appropriate for a multi-user site. TrustAllContent: false # Cache parameters for WebDAV content serving: WebDAVCache: # Time to cache manifests, permission checks, and sessions. TTL: 300s # Maximum amount of data cached in /var/cache/arvados/keep. # Can be given as a percentage of filesystem size ("10%") or a # number of bytes ("10 GiB") DiskCacheSize: 10% # Approximate memory limit (in bytes) for session cache. # # Note this applies to the in-memory representation of # projects and collections -- metadata, block locators, # filenames, etc. -- not the file data itself (see # DiskCacheSize). MaxCollectionBytes: 100 MB # Persistent sessions. MaxSessions: 100 # Selectively set permissions for regular users and admins to # download or upload data files using the upload/download # features for Workbench, WebDAV and S3 API support. WebDAVPermission: User: Download: true Upload: true Admin: Download: true Upload: true # Selectively set permissions for regular users and admins to be # able to download or upload blocks using arv-put and # arv-get from outside the cluster. KeepproxyPermission: User: Download: true Upload: true Admin: Download: true Upload: true # Post upload / download events to the API server logs table, so # that they can be included in the arv-user-activity report. # You can disable this if you find that it is creating excess # load on the API server and you don't need it. WebDAVLogEvents: true # If a client requests partial content past the start of a file, # and a request from the same client for the same file was logged # within the past WebDAVLogDownloadInterval, do not write a new log. # This throttling applies to both printed and API server logs. # This reduces log output when clients like `aws s3 cp` download # one file in small chunks in parallel. # Set this to 0 to disable throttling and log all requests. WebDAVLogDownloadInterval: 30s # Per-connection output buffer for WebDAV downloads. May improve # throughput for large files, particularly when storage volumes # have high latency. # # Size be specified as a number of bytes ("0") or with units # ("128KiB", "1 MB"). WebDAVOutputBuffer: 0 Login: # One of the following mechanisms (Google, PAM, LDAP, or # LoginCluster) should be enabled; see # https://doc.arvados.org/install/setup-login.html Google: # Authenticate with Google. Enable: false # Use the Google Cloud console to enable the People API (APIs # and Services > Enable APIs and services > Google People API # > Enable), generate a Client ID and secret (APIs and # Services > Credentials > Create credentials > OAuth client # ID > Web application) and add your controller's /login URL # (e.g., "https://zzzzz.example.com/login") as an authorized # redirect URL. ClientID: "" ClientSecret: "" # Allow users to log in to existing accounts using any verified # email address listed by their Google account. If true, the # Google People API must be enabled in order for Google login to # work. If false, only the primary email address will be used. AlternateEmailAddresses: true # Send additional parameters with authentication requests. See # https://developers.google.com/identity/protocols/oauth2/openid-connect#authenticationuriparameters # for a list of supported parameters. AuthenticationRequestParameters: # Show the "choose which Google account" page, even if the # client is currently logged in to exactly one Google # account. prompt: select_account SAMPLE: "" OpenIDConnect: # Authenticate with an OpenID Connect provider. Enable: false # Issuer URL, e.g., "https://login.example.com". # # This must be exactly equal to the URL returned by the issuer # itself in its config response ("isser" key). If the # configured value is "https://example" and the provider # returns "https://example:443" or "https://example/" then # login will fail, even though those URLs are equivalent (RFC # 3986). Issuer: "" # Your client ID and client secret (supplied by the provider). ClientID: "" ClientSecret: "" # OpenID claim field containing the user's email # address. Normally "email"; see # https://openid.net/specs/openid-connect-core-1_0.html#StandardClaims EmailClaim: "email" # OpenID claim field containing the email verification # flag. Normally "email_verified". To accept every returned # email address without checking a "verified" field at all, # use an empty string "". EmailVerifiedClaim: "email_verified" # OpenID claim field containing the user's preferred # username. If empty, use the mailbox part of the user's email # address. UsernameClaim: "" # Send additional parameters with authentication requests, # like {display: page, prompt: consent}. See # https://openid.net/specs/openid-connect-core-1_0.html#AuthRequest # and refer to your provider's documentation for supported # parameters. AuthenticationRequestParameters: SAMPLE: "" # Accept an OIDC access token as an API token if the OIDC # provider's UserInfo endpoint accepts it. # # AcceptAccessTokenScope should also be used when enabling # this feature. AcceptAccessToken: false # Before accepting an OIDC access token as an API token, first # check that it is a JWT whose "scope" value includes this # value. Example: "https://zzzzz.example.com/" (your Arvados # API endpoint). # # If this value is empty and AcceptAccessToken is true, all # access tokens will be accepted regardless of scope, # including non-JWT tokens. This is not recommended. AcceptAccessTokenScope: "" PAM: # Use PAM to authenticate users. Enable: false # PAM service name. PAM will apply the policy in the # corresponding config file (e.g., /etc/pam.d/arvados) or, if # there is none, the default "other" config. Service: arvados # Domain name (e.g., "example.com") to use to construct the # user's email address if PAM authentication returns a # username with no "@". If empty, use the PAM username as the # user's email address, whether or not it contains "@". # # Note that the email address is used as the primary key for # user records when logging in. Therefore, if you change # PAMDefaultEmailDomain after the initial installation, you # should also update existing user records to reflect the new # domain. Otherwise, next time those users log in, they will # be given new accounts instead of accessing their existing # accounts. DefaultEmailDomain: "" LDAP: # Use an LDAP service to authenticate users. Enable: false # Server URL, like "ldap://ldapserver.example.com:389" or # "ldaps://ldapserver.example.com:636". URL: "ldap://ldap:389" # Use StartTLS upon connecting to the server. StartTLS: true # Skip TLS certificate name verification. InsecureTLS: false # Mininum TLS version to negotiate when connecting to server # (ldaps://... or StartTLS). It may be necessary to set this # to "1.1" for compatibility with older LDAP servers that fail # with 'LDAP Result Code 200 "Network Error": TLS handshake # failed (tls: server selected unsupported protocol version # 301)'. # # If blank, use the recommended minimum version (1.2). MinTLSVersion: "" # Strip the @domain part if a user supplies an email-style # username with this domain. If "*", strip any user-provided # domain. If "", never strip the domain part. Example: # "example.com" StripDomain: "" # If, after applying StripDomain, the username contains no "@" # character, append this domain to form an email-style # username. Example: "example.com" AppendDomain: "" # The LDAP attribute to filter on when looking up a username # (after applying StripDomain and AppendDomain). SearchAttribute: uid # Bind with this username (DN or UPN) and password when # looking up the user record. # # Example user: "cn=admin,dc=example,dc=com" SearchBindUser: "" SearchBindPassword: "" # Directory base for username lookup. Example: # "ou=Users,dc=example,dc=com" SearchBase: "" # Additional filters to apply when looking up users' LDAP # entries. This can be used to restrict access to a subset of # LDAP users, or to disambiguate users from other directory # entries that have the SearchAttribute present. # # Special characters in assertion values must be escaped (see # RFC 4515). # # Example: "(objectClass=person)" SearchFilters: "" # LDAP attribute to use as the user's email address. # # Important: This must not be an attribute whose value can be # edited in the directory by the users themselves. Otherwise, # users can take over other users' Arvados accounts trivially # (email address is the primary key for Arvados accounts.) EmailAttribute: mail # LDAP attribute to use as the preferred Arvados username. If # no value is found (or this config is empty) the username # originally supplied by the user will be used. UsernameAttribute: uid Test: # Authenticate users listed here in the config file. This # feature is intended to be used in test environments, and # should not be used in production. Enable: false Users: SAMPLE: Email: alice@example.com Password: xyzzy # The cluster ID to delegate the user database. When set, # logins on this cluster will be redirected to the login cluster # (login cluster must appear in RemoteClusters with Proxy: true) LoginCluster: "" # How long a cached token belonging to a remote cluster will # remain valid before it needs to be revalidated. RemoteTokenRefresh: 5m # How long a client token created from a login flow will be valid without # asking the user to re-login. Example values: 60m, 8h. # Default value zero means tokens don't have expiration. TokenLifetime: 0s # If true (default), tokens are allowed to create new tokens and # view existing tokens belonging to the same user. # If false, tokens are not allowed to view or create other # tokens. New tokens can only be created by going through login # again. IssueTrustedTokens: true # Origins (scheme://host[:port]) of clients trusted to receive # new tokens via login process. The ExternalURLs of the local # Workbench1 and Workbench2 are trusted implicitly and do not # need to be listed here. If this is a LoginCluster, you # probably want to include the other Workbench instances in the # federation in this list. # # A wildcard like "https://*.example" will match client URLs # like "https://a.example" and "https://a.b.c.example". # # Example: # # TrustedClients: # "https://workbench.other-cluster.example": {} # "https://workbench2.other-cluster.example": {} TrustedClients: SAMPLE: {} # Treat any origin whose host part is "localhost" or a private # IP address (e.g., http://10.0.0.123:3000/) as if it were # listed in TrustedClients. # # Intended only for test/development use. Not appropriate for # production use. TrustPrivateNetworks: false TLS: # Use "file:///var/lib/acme/live/example.com/cert" and # ".../privkey" to load externally managed certificates. Certificate: "" Key: "" # Accept invalid certificates when connecting to servers. Never # use this in production. Insecure: false ACME: # Obtain certificates automatically for ExternalURL domains # using an ACME server and http-01 validation. # # To use Let's Encrypt, specify "LE". To use the Let's # Encrypt staging environment, specify "LE-staging". To use a # different ACME server, specify the full directory URL # ("https://..."). # # Note: this feature is not yet implemented in released # versions, only in the alpha/prerelease arvados-server-easy # package. # # Implies agreement with the server's terms of service. Server: "" Containers: # List of supported Docker Registry image formats that compute nodes # are able to use. `arv keep docker` will error out if a user tries # to store an image with an unsupported format. Use an empty array # to skip the compatibility check (and display a warning message to # that effect). # # Example for sites running docker < 1.10: {"v1": {}} # Example for sites running docker >= 1.10: {"v2": {}} # Example for disabling check: {} SupportedDockerImageFormats: "v2": {} SAMPLE: {} # Include details about job reuse decisions in the server log. This # causes additional database queries to run, so it should not be # enabled unless you expect to examine the resulting logs for # troubleshooting purposes. LogReuseDecisions: false # Default value for keep_cache_ram of a container's # runtime_constraints. Note: this gets added to the RAM request # used to allocate a VM or submit an HPC job. # # If this is zero, container requests that don't specify RAM or # disk cache size will use a disk cache, sized to the # container's RAM requirement (but with minimum 2 GiB and # maximum 32 GiB). # # Note: If you change this value, containers that used the previous # default value will only be reused by container requests that # explicitly specify the previous value in their keep_cache_ram # runtime constraint. DefaultKeepCacheRAM: 0 # Number of times a container can be unlocked before being # automatically cancelled. MaxDispatchAttempts: 10 # Default value for container_count_max for container requests. This is the # number of times Arvados will create a new container to satisfy a container # request. If a container is cancelled it will retry a new container if # container_count < container_count_max on any container requests associated # with the cancelled container. MaxRetryAttempts: 3 # Schedule all child containers on preemptible instances (e.g. AWS # Spot Instances) even if not requested by the submitter. # # If false, containers are scheduled on preemptible instances # only when requested by the submitter. # # This flag is ignored if no preemptible instance types are # configured, and has no effect on top-level containers. AlwaysUsePreemptibleInstances: false # Automatically add a preemptible variant for every # non-preemptible entry in InstanceTypes below. The maximum bid # price for the preemptible variant will be the non-preemptible # price multiplied by PreemptiblePriceFactor. If 0, preemptible # variants are not added automatically. # # A price factor of 1.0 is a reasonable starting point. PreemptiblePriceFactor: 0 # When the lowest-priced instance type for a given container is # not available, try other instance types, up to the indicated # maximum price factor. # # For example, with AvailabilityPriceFactor 1.5, if the # lowest-cost instance type A suitable for a given container # costs $2/h, Arvados may run the container on any instance type # B costing $3/h or less when instance type A is not available # or an idle instance of type B is already running. MaximumPriceFactor: 1.5 # PEM encoded SSH key (RSA, DSA, or ECDSA) used by the # cloud dispatcher for executing containers on worker VMs. # Begins with "-----BEGIN RSA PRIVATE KEY-----\n" # and ends with "\n-----END RSA PRIVATE KEY-----\n". # # Use "file:///absolute/path/to/key" to load the key from a # separate file instead of embedding it in the configuration # file. DispatchPrivateKey: "" # Maximum time to wait for workers to come up before abandoning # stale locks from a previous dispatch process. StaleLockTimeout: 1m # The crunch-run command used to start a container on a worker node. # # When dispatching to cloud VMs, this is used only if # DeployRunnerBinary in the CloudVMs section is set to the empty # string. CrunchRunCommand: "crunch-run" # Extra arguments to add to crunch-run invocation # Example: ["--cgroup-parent-subsystem=memory"] CrunchRunArgumentsList: [] # Extra RAM to reserve on the node, in addition to # the amount specified in the container's RuntimeConstraints ReserveExtraRAM: 550MiB # Minimum time between two attempts to run the same container MinRetryPeriod: 0s # Container runtime: "docker" (default) or "singularity" RuntimeEngine: docker # When running a container, run a dedicated keepstore process, # using the specified number of 64 MiB memory buffers per # allocated CPU core (VCPUs in the container's runtime # constraints). The dedicated keepstore handles I/O for # collections mounted in the container, as well as saving # container logs. # # A zero value disables this feature. # # In order for this feature to be activated, no volume may use # AccessViaHosts, and no writable volume may have Replication # lower than Collections.DefaultReplication. If these # requirements are not satisfied, the feature is disabled # automatically regardless of the value given here. # # When an HPC dispatcher is in use (see SLURM and LSF sections), # this feature depends on the operator to ensure an up-to-date # cluster configuration file (/etc/arvados/config.yml) is # available on all compute nodes. If it is missing or not # readable by the crunch-run user, the feature will be disabled # automatically. To read it from a different location, add a # "-config=/path/to/config.yml" argument to # CrunchRunArgumentsList above. # # When the cloud dispatcher is in use (see CloudVMs section) and # this configuration is enabled, the entire cluster # configuration file, including the system root token, is copied # to the worker node and held in memory for the duration of the # container. LocalKeepBlobBuffersPerVCPU: 1 # When running a dedicated keepstore process for a container # (see LocalKeepBlobBuffersPerVCPU), write keepstore log # messages to keepstore.txt in the container's log collection. # # These log messages can reveal some volume configuration # details, error messages from the cloud storage provider, etc., # which are not otherwise visible to users. # # Accepted values: # * "none" -- no keepstore.txt file # * "all" -- all logs, including request and response lines # * "errors" -- all logs except "response" logs with 2xx # response codes and "request" logs LocalKeepLogsToContainerLog: none Logging: # Container logs are written to Keep and saved in a # collection, which is updated periodically while the # container runs. This value sets the interval between # collection updates. LogUpdatePeriod: 30m # The log collection is also updated when the specified amount of # log data (given in bytes) is produced in less than one update # period. LogUpdateSize: 32MiB ShellAccess: # An admin user can use "arvados-client shell" to start an # interactive shell (with any user ID) in any running # container. Admin: false # Any user can use "arvados-client shell" to start an # interactive shell (with any user ID) in any running # container that they started, provided it isn't also # associated with a different user's container request. # # Interactive sessions make it easy to alter the container's # runtime environment in ways that aren't recorded or # reproducible. Consider the implications for automatic # container reuse before enabling and using this feature. In # particular, note that starting an interactive session does # not disqualify a container from being reused by a different # user/workflow in the future. User: false SLURM: PrioritySpread: 0 SbatchArgumentsList: [] SbatchEnvironmentVariables: SAMPLE: "" LSF: # Arguments to bsub when submitting Arvados containers as LSF jobs. # # Template variables starting with % will be substituted as follows: # # %U uuid # %C number of VCPUs # %M memory in MB # %T tmp in MB # %G number of GPU devices (runtime_constraints.cuda.device_count) # %W maximum run time in minutes (see MaxRunTimeOverhead and # MaxRunTimeDefault below) # # Use %% to express a literal %. For example, the %%J in the # default argument list will be changed to %J, which is # interpreted by bsub itself. # # Note that the default arguments cause LSF to write two files # in /tmp on the compute node each time an Arvados container # runs. Ensure you have something in place to delete old files # from /tmp, or adjust the "-o" and "-e" arguments accordingly. # # If ["-We", "%W"] or ["-W", "%W"] appear in this argument # list, and MaxRunTimeDefault is not set (see below), both of # those arguments will be dropped from the argument list when # running a container that has no max_run_time value. BsubArgumentsList: ["-o", "/tmp/crunch-run.%%J.out", "-e", "/tmp/crunch-run.%%J.err", "-J", "%U", "-n", "%C", "-D", "%MMB", "-R", "rusage[mem=%MMB:tmp=%TMB] span[hosts=1]", "-R", "select[mem>=%MMB]", "-R", "select[tmp>=%TMB]", "-R", "select[ncpus>=%C]", "-We", "%W"] # Arguments that will be appended to the bsub command line # when submitting Arvados containers as LSF jobs with # runtime_constraints.cuda.device_count > 0 BsubCUDAArguments: ["-gpu", "num=%G"] # Use sudo to switch to this user account when submitting LSF # jobs. # # This account must exist on the hosts where LSF jobs run # ("execution hosts"), as well as on the host where the # Arvados LSF dispatcher runs ("submission host"). BsubSudoUser: "crunch" # When passing the scheduling_constraints.max_run_time value # to LSF via "%W", add this much time to account for # crunch-run startup/shutdown overhead. MaxRunTimeOverhead: 5m # If non-zero, MaxRunTimeDefault is used as the default value # for max_run_time for containers that do not specify a time # limit. MaxRunTimeOverhead will be added to this. # # Example: # MaxRunTimeDefault: 2h MaxRunTimeDefault: 0 CloudVMs: # Enable the cloud scheduler. Enable: false # Name/number of port where workers' SSH services listen. SSHPort: "22" # Interval between queue polls. PollInterval: 10s # Shell command to execute on each worker to determine whether # the worker is booted and ready to run containers. It should # exit zero if the worker is ready. BootProbeCommand: "systemctl is-system-running" # Minimum interval between consecutive probes to a single # worker. ProbeInterval: 10s # Maximum probes per second, across all workers in a pool. MaxProbesPerSecond: 10 # Time before repeating SIGTERM when killing a container. TimeoutSignal: 5s # Time to give up on a process (most likely arv-mount) that # still holds a container lockfile after its main supervisor # process has exited, and declare the instance broken. TimeoutStaleRunLock: 5s # Time to give up on SIGTERM and write off the worker. TimeoutTERM: 2m # Maximum create/destroy-instance operations per second (0 = # unlimited). MaxCloudOpsPerSecond: 10 # Maximum concurrent instance creation operations (0 = unlimited). # # MaxConcurrentInstanceCreateOps limits the number of instance creation # requests that can be in flight at any one time, whereas # MaxCloudOpsPerSecond limits the number of create/destroy operations # that can be started per second. # # Because the API for instance creation on Azure is synchronous, it is # recommended to increase MaxConcurrentInstanceCreateOps when running # on Azure. When using managed images, a value of 20 would be # appropriate. When using Azure Shared Image Galeries, it could be set # higher. For more information, see # https://docs.microsoft.com/en-us/azure/virtual-machines/linux/capture-image # # MaxConcurrentInstanceCreateOps can be increased for other cloud # providers too, if desired. MaxConcurrentInstanceCreateOps: 1 # The maximum number of instances to run at a time, or 0 for # unlimited. # # If more instances than this are already running and busy # when the dispatcher starts up, the running containers will # be allowed to finish before the excess instances are shut # down. MaxInstances: 64 # The minimum number of instances expected to be runnable # without reaching a provider-imposed quota. # # This is used as the initial value for the dispatcher's # dynamic instance limit, which increases (up to MaxInstances) # as containers start up successfully and decreases in # response to high API load and cloud quota errors. # # Setting this to 0 means the dynamic instance limit will # start at MaxInstances. # # Situations where you may want to set this (to a value less # than MaxInstances) would be when there is significant # variability or uncertainty in the actual cloud resources # available. Upon reaching InitialQuotaEstimate the # dispatcher will switch to a more conservative behavior with # slower instance start to avoid over-shooting cloud resource # limits. InitialQuotaEstimate: 0 # Maximum fraction of available instance capacity allowed to # run "supervisor" containers at any given time. A supervisor # is a container whose purpose is mainly to submit and manage # other containers, such as arvados-cwl-runner workflow # runner. # # If there is a hard limit on the amount of concurrent # containers that the cluster can run, it is important to # avoid crowding out the containers doing useful work with # containers who just create more work. # # For example, with the default MaxInstances of 64, it will # schedule at most floor(64*0.50) = 32 concurrent workflow # runners, ensuring 32 slots are available for work. SupervisorFraction: 0.50 # Interval between cloud provider syncs/updates ("list all # instances"). SyncInterval: 1m # Time to leave an idle worker running (in case new containers # appear in the queue that it can run) before shutting it # down. TimeoutIdle: 1m # Time to wait for a new worker to boot (i.e., pass # BootProbeCommand) before giving up and shutting it down. TimeoutBooting: 10m # Maximum time a worker can stay alive with no successful # probes before being automatically shut down. TimeoutProbe: 10m # Time after shutting down a worker to retry the # shutdown/destroy operation. TimeoutShutdown: 10s # Worker VM image ID. # (aws) AMI identifier # (azure) managed disks: the name of the managed disk image # (azure) shared image gallery: the name of the image definition. Also # see the SharedImageGalleryName and SharedImageGalleryImageVersion fields. # (azure) unmanaged disks (deprecated): the complete URI of the VHD, e.g. # https://xxxxx.blob.core.windows.net/system/Microsoft.Compute/Images/images/xxxxx.vhd ImageID: "" # Shell script to run on new instances using the cloud # provider's UserData (EC2) or CustomData (Azure) feature. # # It is not necessary to include a #!/bin/sh line. InstanceInitCommand: "" # An executable file (located on the dispatcher host) to be # copied to cloud instances at runtime and used as the # "crunch-run" container runner/supervisor. The default value # is the dispatcher program itself. # # Use an empty string to disable this step: nothing will be # copied, and cloud instances are assumed to have a suitable # version of crunch-run installed; see CrunchRunCommand above. DeployRunnerBinary: "/proc/self/exe" # Directory to store the crunch-run binary on cloud instances # (see DeployRunnerBinary above). The "mkdir -p" command will # be used to create the directory and its parents if needed. DeployRunnerDirectory: /tmp/arvados-crunch-run # Install the Dispatcher's SSH public key (derived from # DispatchPrivateKey) when creating new cloud # instances. Change this to false if you are using a different # mechanism to pre-install the public key on new instances. DeployPublicKey: true # Tags to add on all resources (VMs, NICs, disks) created by # the container dispatcher. (Arvados's own tags -- # InstanceType, IdleBehavior, and InstanceSecret -- will also # be added.) ResourceTags: SAMPLE: "tag value" # Prefix for predefined tags used by Arvados (InstanceSetID, # InstanceType, InstanceSecret, IdleBehavior). With the # default value "Arvados", tags are "ArvadosInstanceSetID", # "ArvadosInstanceSecret", etc. # # This should only be changed while no cloud resources are in # use and the cloud dispatcher is not running. Otherwise, # VMs/resources that were added using the old tag prefix will # need to be detected and cleaned up manually. TagKeyPrefix: Arvados # Cloud driver: "azure" (Microsoft Azure), "ec2" (Amazon AWS), # or "loopback" (run containers on dispatch host for testing # purposes). Driver: ec2 # Cloud-specific driver parameters. DriverParameters: # (ec2) Credentials. Omit or leave blank if using IAM role. AccessKeyID: "" SecretAccessKey: "" # (ec2) Instance configuration. # (ec2) Region, like "us-east-1". Region: "" # (ec2) Security group IDs. Omit or use {} to use the # default security group. SecurityGroupIDs: "SAMPLE": {} # (ec2) One or more subnet IDs. Omit or leave empty to let # AWS choose a default subnet from your default VPC. If # multiple subnets are configured here (enclosed in brackets # like [subnet-abc123, subnet-def456]) the cloud dispatcher # will detect subnet-related errors and retry using a # different subnet. Most sites specify one subnet. SubnetID: "" EBSVolumeType: gp2 AdminUsername: debian # (ec2) name of the IAMInstanceProfile for instances started by # the cloud dispatcher. Leave blank when not needed. IAMInstanceProfile: "" # (ec2) how often to look up spot instance pricing data # (only while running spot instances) for the purpose of # calculating container cost estimates. A value of 0 # disables spot price lookups entirely. SpotPriceUpdateInterval: 24h # (ec2) per-GiB-month cost of EBS volumes. Matches # EBSVolumeType. Used to account for AddedScratch when # calculating container cost estimates. Note that # https://aws.amazon.com/ebs/pricing/ defines GB to mean # GiB, so an advertised price $0.10/GB indicates a real # price of $0.10/GiB and can be entered here as 0.10. EBSPrice: 0.10 # (ec2) Mapping of alphabetic instance type prefix to # instance quota group. Any prefix not listed here will be # treated as a distinct instance quota group. For example, # "trn1.2xlarge" will implicitly belong to instance quota # group "trn". # # Knowing that multiple instance types belong to the same # quota group enables the dispatcher to minimize futile # attempts to create new instances when a quota has been # reached. # # All keys must be lowercase. InstanceTypeQuotaGroups: a: standard c: standard d: standard h: standard i: standard m: standard r: standard t: standard z: standard vt: g p5: p5 # (azure) Credentials. SubscriptionID: "" ClientID: "" ClientSecret: "" TenantID: "" # (azure) Instance configuration. CloudEnvironment: AzurePublicCloud Location: centralus # (azure) The resource group where the VM and virtual NIC will be # created. ResourceGroup: "" # (azure) The resource group of the Network to use for the virtual # NIC (if different from ResourceGroup) NetworkResourceGroup: "" Network: "" Subnet: "" # (azure) managed disks: The resource group where the managed disk # image can be found (if different from ResourceGroup). ImageResourceGroup: "" # (azure) shared image gallery: the name of the gallery SharedImageGalleryName: "" # (azure) shared image gallery: the version of the image definition SharedImageGalleryImageVersion: "" # (azure) unmanaged disks (deprecated): Where to store the VM VHD blobs StorageAccount: "" BlobContainer: "" # (azure) How long to wait before deleting VHD and NIC # objects that are no longer being used. DeleteDanglingResourcesAfter: 20s # Account (that already exists in the VM image) that will be # set up with an ssh authorized key to allow the compute # dispatcher to connect. AdminUsername: arvados InstanceTypes: # Use the instance type name as the key (in place of "SAMPLE" in # this sample entry). SAMPLE: # Cloud provider's instance type. Defaults to the configured type name. ProviderType: "" VCPUs: 1 RAM: 128MiB IncludedScratch: 16GB AddedScratch: 0 # Hourly price ($), used to select node types for containers, # and to calculate estimated container costs. For spot # instances on EC2, this is also used as the maximum price # when launching spot instances, while the estimated container # cost is computed based on the current spot price according # to AWS. On Azure, and on-demand instances on EC2, the price # given here is used to compute container cost estimates. Price: 0.1 Preemptible: false # Include this section if the node type includes GPU (CUDA) support CUDA: DriverVersion: "11.0" HardwareCapability: "9.0" DeviceCount: 1 StorageClasses: # If you use multiple storage classes, specify them here, using # the storage class name as the key (in place of "SAMPLE" in # this sample entry). # # Further info/examples: # https://doc.arvados.org/admin/storage-classes.html SAMPLE: # Priority determines the order volumes should be searched # when reading data, in cases where a keepstore server has # access to multiple volumes with different storage classes. Priority: 0 # Default determines which storage class(es) should be used # when a user/client writes data or saves a new collection # without specifying storage classes. # # If any StorageClasses are configured, at least one of them # must have Default: true. Default: true Volumes: SAMPLE: # AccessViaHosts specifies which keepstore processes can read # and write data on the volume. # # For a local filesystem, AccessViaHosts has one entry, # indicating which server the filesystem is located on. # # For a network-attached backend accessible by all keepstore # servers, like a cloud storage bucket or an NFS mount, # AccessViaHosts can be empty/omitted. # # Further info/examples: # https://doc.arvados.org/install/configure-fs-storage.html # https://doc.arvados.org/install/configure-s3-object-storage.html # https://doc.arvados.org/install/configure-azure-blob-storage.html AccessViaHosts: SAMPLE: ReadOnly: false "http://host1.example:25107": {} ReadOnly: false # AllowTrashWhenReadOnly enables unused and overreplicated # blocks to be trashed/deleted even when ReadOnly is # true. Normally, this is false and ReadOnly prevents all # trash/delete operations as well as writes. AllowTrashWhenReadOnly: false Replication: 1 StorageClasses: # If you have configured storage classes (see StorageClasses # section above), add an entry here for each storage class # satisfied by this volume. SAMPLE: true Driver: S3 DriverParameters: # for s3 driver -- see # https://doc.arvados.org/install/configure-s3-object-storage.html AccessKeyID: aaaaa SecretAccessKey: aaaaa Endpoint: "" Region: us-east-1 Bucket: aaaaa LocationConstraint: false V2Signature: false UsePathStyle: false IndexPageSize: 1000 ConnectTimeout: 1m ReadTimeout: 10m RaceWindow: 24h PrefixLength: 0 # For S3 driver, potentially unsafe tuning parameter, # intentionally excluded from main documentation. # # Enable deletion (garbage collection) even when the # configured BlobTrashLifetime is zero. WARNING: eventual # consistency may result in race conditions that can cause # data loss. Do not enable this unless you understand and # accept the risk. UnsafeDelete: false # for azure driver -- see # https://doc.arvados.org/install/configure-azure-blob-storage.html StorageAccountName: aaaaa StorageAccountKey: aaaaa StorageBaseURL: core.windows.net ContainerName: aaaaa RequestTimeout: 30s ListBlobsRetryDelay: 10s ListBlobsMaxAttempts: 10 MaxGetBytes: 0 WriteRaceInterval: 15s WriteRacePollTime: 1s # for local directory driver -- see # https://doc.arvados.org/install/configure-fs-storage.html Root: /var/lib/arvados/keep-data # For local directory driver, potentially confusing tuning # parameter, intentionally excluded from main documentation. # # When true, read and write operations (for whole 64MiB # blocks) on an individual volume will queued and issued # serially. When false, read and write operations will be # issued concurrently. # # May possibly improve throughput if you have physical spinning disks # and experience contention when there are multiple requests # to the same volume. # # Otherwise, when using SSDs, RAID, or a shared network filesystem, you # should leave this alone. Serialize: false RemoteClusters: "*": Host: "" Proxy: false Scheme: https Insecure: false ActivateUsers: false SAMPLE: # API endpoint host or host:port; default is {id}.arvadosapi.com Host: sample.arvadosapi.com # Perform a proxy request when a local client requests an # object belonging to this remote. Proxy: false # Default "https". Can be set to "http" for testing. Scheme: https # Disable TLS verify. Can be set to true for testing. Insecure: false # When users present tokens issued by this remote cluster, and # their accounts are active on the remote cluster, activate # them on this cluster too. ActivateUsers: false Workbench: # Workbench1 configs Theme: default ActivationContactLink: mailto:info@arvados.org ArvadosDocsite: https://doc.arvados.org ArvadosPublicDataDocURL: https://playground.arvados.org/projects/public ShowUserAgreementInline: false # Set this configuration to true to avoid providing an easy way for users # to share data with unauthenticated users; this may be necessary on # installations where strict data access controls are needed. DisableSharingURLsUI: false # Below is a sample setting of user_profile_form_fields config parameter. # This configuration parameter should be set to either false (to disable) or # to a map as shown below. # Configure the map of input fields to be displayed in the profile page # using the attribute "key" for each of the input fields. # This sample shows configuration with one required and one optional form fields. # For each of these input fields: # You can specify "Type" as "text" or "select". # List the "Options" to be displayed for each of the "select" menu. # Set "Required" as "true" for any of these fields to make them required. # If any of the required fields are missing in the user's profile, the user will be # redirected to the profile page before they can access any Workbench features. UserProfileFormFields: SAMPLE: Type: select FormFieldTitle: Best color FormFieldDescription: your favorite color Required: false Position: 1 Options: red: {} blue: {} green: {} SAMPLE: {} # exampleTextValue: # key that will be set in properties # Type: text # # FormFieldTitle: "" # FormFieldDescription: "" # Required: true # Position: 1 # exampleOptionsValue: # Type: select # FormFieldTitle: "" # FormFieldDescription: "" # Required: true # Position: 1 # Options: # red: {} # blue: {} # yellow: {} # Use "UserProfileFormMessage to configure the message you want # to display on the profile page. UserProfileFormMessage: 'Welcome to Arvados. All <span style="color:red">required fields</span> must be completed before you can proceed.' SiteName: Arvados Workbench # Workbench2 configs FileViewersConfigURL: "" # Idle time after which the user's session will be auto closed. # This feature is disabled when set to zero. IdleTimeout: 0s # UUID of a collection. This collection should be shared with # all users. Workbench will look for a file "banner.html" in # this collection and display its contents (should be # HTML-formatted text) when users first log in to Workbench. BannerUUID: "" # Workbench welcome screen, this is HTML text that will be # incorporated directly onto the page. WelcomePageHTML: | <img src="/arvados-logo-big.png" style="width: 20%; float: right; padding: 1em;" /> <h2>Please log in.</h2> <p>If you have never used Arvados Workbench before, logging in for the first time will automatically create a new account.</p> <i>Arvados Workbench uses your information only for identification, and does not retrieve any other personal information.</i> # Workbench screen displayed to inactive users. This is HTML # text that will be incorporated directly onto the page. InactivePageHTML: | <img src="/arvados-logo-big.png" style="width: 20%; float: right; padding: 1em;" /> <h3>Hi! You're logged in, but...</h3> <p>Your account is inactive.</p> <p>An administrator must activate your account before you can get any further.</p> # Connecting to Arvados shell VMs tends to be site-specific. # Put any special instructions here. This is HTML text that will # be incorporated directly onto the Workbench page. SSHHelpPageHTML: | <a href="https://doc.arvados.org/user/getting_started/ssh-access-unix.html">Accessing an Arvados VM with SSH</a> (generic instructions). Site configurations vary. Contact your local cluster administrator if you have difficulty accessing an Arvados shell node. # Sample text if you are using a "switchyard" ssh proxy. # Replace "zzzzz" with your Cluster ID. #SSHHelpPageHTML: | # <p>Add a section like this to your SSH configuration file ( <i>~/.ssh/config</i>):</p> # <pre>Host *.zzzzz # TCPKeepAlive yes # ServerAliveInterval 60 # ProxyCommand ssh -p2222 turnout@switchyard.zzzzz.arvadosapi.com -x -a $SSH_PROXY_FLAGS %h # </pre> # If you are using a switchyard ssh proxy, shell node hostnames # may require a special hostname suffix. In the sample ssh # configuration above, this would be ".zzzzz" # This is added to the hostname in the "command line" column # the Workbench "shell VMs" page. # # If your shell nodes are directly accessible by users without a # proxy and have fully qualified host names, you should leave # this blank. SSHHelpHostSuffix: "" # (Experimental) Restart services automatically when config file # changes are detected. Only supported by `arvados-server boot` in # dev/test mode. AutoReloadConfig: false
The content of this documentation is licensed under the
Creative
Commons Attribution-Share Alike 3.0 United States licence.
Code samples in this documentation are licensed under the
Apache License, Version 2.0.