1 # Google Compute Engine configuration for Arvados Node Manager.
2 # All times are in seconds unless specified otherwise.
5 # Node Manager will ensure that there are at least this many nodes
6 # running at all times.
9 # Node Manager will not start any compute nodes when at least this
13 # Poll compute nodes and Arvados for new information every N seconds.
16 # Polls have exponential backoff when services fail to respond.
17 # This is the longest time to wait between polls.
20 # If Node Manager can't succesfully poll a service for this long,
21 # it will never start or stop compute nodes, on the assumption that its
22 # information is too outdated.
23 poll_stale_after = 600
25 # "Node stale time" affects two related behaviors.
26 # 1. If a compute node has been running for at least this long, but it
27 # isn't paired with an Arvados node, do not shut it down, but leave it alone.
28 # This prevents the node manager from shutting down a node that might
29 # actually be doing work, but is having temporary trouble contacting the
31 # 2. When the Node Manager starts a new compute node, it will try to reuse
32 # an Arvados node that hasn't been updated for this long.
33 node_stale_after = 14400
35 # File path for Certificate Authorities
36 certs_file = /etc/ssl/certs/ca-certificates.crt
40 file = /var/log/arvados/node-manager.log
42 # Log level for most Node Manager messages.
43 # Choose one of DEBUG, INFO, WARNING, ERROR, or CRITICAL.
44 # WARNING lets you know when polling a service fails.
45 # INFO additionally lets you know when a compute node is started or stopped.
48 # You can also set different log levels for specific libraries.
49 # Pykka is the Node Manager's actor library.
50 # Setting this to DEBUG will display tracebacks for uncaught
51 # exceptions in the actors, but it's also very chatty.
54 # Setting apiclient to INFO will log the URL of every Arvados API request.
58 host = zyxwv.arvadosapi.com
62 # Accept an untrusted SSL certificate from the API server?
68 # Shutdown windows define periods of time when a node may and may not
69 # be shut down. These are windows in full minutes, separated by
70 # commas. Counting from the time the node is booted, the node WILL
71 # NOT shut down for N1 minutes; then it MAY shut down for N2 minutes;
72 # then it WILL NOT shut down for N3 minutes; and so on. For example,
73 # "54, 5, 1" means the node may shut down from the 54th to the 59th
74 # minute of each hour of uptime.
75 # GCE bills by the minute, and does not provide information about when
76 # a node booted. Node Manager will store this information in metadata
77 # when it boots a node; if that information is not available, it will
78 # assume the node booted at the epoch. These shutdown settings are
79 # very aggressive. You may want to adjust this if you want more
80 # continuity of service from a single node.
81 shutdown_windows = 20, 999999
84 user_id = client_email_address@developer.gserviceaccount.com
85 key = path_to_certificate.pem
86 project = project-id-from-google-cloud-dashboard
89 # Valid location (zone) names: https://cloud.google.com/compute/docs/zones
90 datacenter = us-central1-a
92 # Optional settings. For full documentation see
93 # http://libcloud.readthedocs.org/en/latest/compute/drivers/gce.html#libcloud.compute.drivers.gce.GCENodeDriver
95 # auth_type = SA # SA, IA or GCE
96 # scopes = https://www.googleapis.com/auth/compute
100 # A comma-separated list of tags that must be applied to a node for it to
101 # be considered a compute node.
102 # The driver will automatically apply these tags to nodes it creates.
103 tags = zyxwv, compute
106 # New compute nodes will send pings to Arvados at this host.
107 # You may specify a port, and use brackets to disambiguate IPv6 addresses.
108 ping_host = hostname:port
110 # A file path for an SSH key that can log in to the compute node.
113 # The GCE image name and network zone name to use when creating new nodes.
115 # network = your_network_name
117 # JSON string of service account authorizations for this cluster.
118 # See http://libcloud.readthedocs.org/en/latest/compute/drivers/gce.html#specifying-service-account-scopes
119 # service_accounts = [{'email':'account@example.com', 'scopes':['storage-ro']}]
122 # You can define any number of Size sections to list node sizes you're
123 # willing to use. The Node Manager should boot the cheapest size(s) that
124 # can run jobs in the queue (N.B.: defining more than one size has not been
127 # The Size fields are interpreted the same way as with a libcloud NodeSize:
128 # http://libcloud.readthedocs.org/en/latest/compute/api.html#libcloud.compute.base.NodeSize
130 # See https://cloud.google.com/compute/docs/machine-types for a list
131 # of known machine types that may be used as a Size parameter.
133 # Each size section MUST define the number of cores are available in this
134 # size class (since libcloud does not provide any consistent API for exposing
136 # You may also want to define the amount of scratch space (expressed
137 # in GB) for Crunch jobs.