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 running at
6 # all times. If node manager needs to start new idle nodes for the purpose of
7 # satisfying min_nodes, it will use the cheapest node type. However, depending
8 # on usage patterns, it may also satisfy min_nodes by keeping alive some
12 # Node Manager will not start any compute nodes when at least this
13 # running at all times. By default, these will be the cheapest node size.
16 # Poll compute nodes and Arvados for new information every N seconds.
19 # Upper limit on rate of spending (in $/hr), will not boot additional nodes
20 # if total price of already running nodes meets or exceeds this threshold.
21 # default 0 means no limit.
24 # Polls have exponential backoff when services fail to respond.
25 # This is the longest time to wait between polls.
28 # If Node Manager can't succesfully poll a service for this long,
29 # it will never start or stop compute nodes, on the assumption that its
30 # information is too outdated.
31 poll_stale_after = 600
33 # "Node stale time" affects two related behaviors.
34 # 1. If a compute node has been running for at least this long, but it
35 # isn't paired with an Arvados node, do not shut it down, but leave it alone.
36 # This prevents the node manager from shutting down a node that might
37 # actually be doing work, but is having temporary trouble contacting the
39 # 2. When the Node Manager starts a new compute node, it will try to reuse
40 # an Arvados node that hasn't been updated for this long.
41 node_stale_after = 14400
43 # Scaling factor to be applied to nodes' available RAM size. Usually there's a
44 # variable discrepancy between the advertised RAM value on cloud nodes and the
45 # actual amount available.
46 # If not set, this value will be set to 0.95
47 node_mem_scaling = 0.95
49 # File path for Certificate Authorities
50 certs_file = /etc/ssl/certs/ca-certificates.crt
54 file = /var/log/arvados/node-manager.log
56 # Log level for most Node Manager messages.
57 # Choose one of DEBUG, INFO, WARNING, ERROR, or CRITICAL.
58 # WARNING lets you know when polling a service fails.
59 # INFO additionally lets you know when a compute node is started or stopped.
62 # You can also set different log levels for specific libraries.
63 # Pykka is the Node Manager's actor library.
64 # Setting this to DEBUG will display tracebacks for uncaught
65 # exceptions in the actors, but it's also very chatty.
68 # Setting apiclient to INFO will log the URL of every Arvados API request.
72 host = zyxwv.arvadosapi.com
76 # Accept an untrusted SSL certificate from the API server?
82 # Shutdown windows define periods of time when a node may and may not
83 # be shut down. These are windows in full minutes, separated by
84 # commas. Counting from the time the node is booted, the node WILL
85 # NOT shut down for N1 minutes; then it MAY shut down for N2 minutes;
86 # then it WILL NOT shut down for N3 minutes; and so on. For example,
87 # "54, 5, 1" means the node may shut down from the 54th to the 59th
88 # minute of each hour of uptime.
89 # GCE bills by the minute, and does not provide information about when
90 # a node booted. Node Manager will store this information in metadata
91 # when it boots a node; if that information is not available, it will
92 # assume the node booted at the epoch. These shutdown settings are
93 # very aggressive. You may want to adjust this if you want more
94 # continuity of service from a single node.
95 shutdown_windows = 20, 999999
98 user_id = client_email_address@developer.gserviceaccount.com
99 key = path_to_certificate.pem
100 project = project-id-from-google-cloud-dashboard
103 # Valid location (zone) names: https://cloud.google.com/compute/docs/zones
104 datacenter = us-central1-a
106 # Optional settings. For full documentation see
107 # http://libcloud.readthedocs.org/en/latest/compute/drivers/gce.html#libcloud.compute.drivers.gce.GCENodeDriver
109 # auth_type = SA # SA, IA or GCE
110 # scopes = https://www.googleapis.com/auth/compute
114 # A comma-separated list of tags that must be applied to a node for it to
115 # be considered a compute node.
116 # The driver will automatically apply these tags to nodes it creates.
117 tags = zyxwv, compute
120 # New compute nodes will send pings to Arvados at this host.
121 # You may specify a port, and use brackets to disambiguate IPv6 addresses.
122 ping_host = hostname:port
124 # A file path for an SSH key that can log in to the compute node.
127 # The GCE image name and network zone name to use when creating new nodes.
129 # network = your_network_name
131 # JSON string of service account authorizations for this cluster.
132 # See http://libcloud.readthedocs.org/en/latest/compute/drivers/gce.html#specifying-service-account-scopes
133 # service_accounts = [{'email':'account@example.com', 'scopes':['storage-ro']}]
136 # You can define any number of Size sections to list node sizes you're
137 # willing to use. The Node Manager should boot the cheapest size(s) that
138 # can run jobs in the queue.
140 # The Size fields are interpreted the same way as with a libcloud NodeSize:
141 # http://libcloud.readthedocs.org/en/latest/compute/api.html#libcloud.compute.base.NodeSize
143 # See https://cloud.google.com/compute/docs/machine-types for a list
144 # of known machine types that may be used as a Size parameter.
146 # Each size section MUST define the number of cores are available in this
147 # size class (since libcloud does not provide any consistent API for exposing
149 # You may also want to define the amount of scratch space (expressed
150 # in GB) for Crunch jobs.
151 # You can also override Google's provided data fields (such as price per hour)
152 # by setting them here.