1 # Azure configuration for Arvados Node Manager.
2 # All times are in seconds unless specified otherwise.
5 # The management server responds to http://addr:port/status.json with
6 # a snapshot of internal state.
8 # Management server listening address (default 127.0.0.1)
11 # Management server port number (default -1, server is disabled)
15 # The dispatcher can customize the start and stop procedure for
16 # cloud nodes. For example, the SLURM dispatcher drains nodes
17 # through SLURM before shutting them down.
20 # Node Manager will ensure that there are at least this many nodes running at
21 # all times. If node manager needs to start new idle nodes for the purpose of
22 # satisfying min_nodes, it will use the cheapest node type. However, depending
23 # on usage patterns, it may also satisfy min_nodes by keeping alive some
24 # more-expensive nodes
27 # Node Manager will not start any compute nodes when at least this
31 # Upper limit on rate of spending (in $/hr), will not boot additional nodes
32 # if total price of already running nodes meets or exceeds this threshold.
33 # default 0 means no limit.
36 # Poll Azure nodes and Arvados for new information every N seconds.
39 # Polls have exponential backoff when services fail to respond.
40 # This is the longest time to wait between polls.
43 # If Node Manager can't succesfully poll a service for this long,
44 # it will never start or stop compute nodes, on the assumption that its
45 # information is too outdated.
46 poll_stale_after = 600
48 # If Node Manager boots a cloud node, and it does not pair with an Arvados
49 # node before this long, assume that there was a cloud bootstrap failure and
50 # shut it down. Note that normal shutdown windows apply (see the Cloud
51 # section), so this should be shorter than the first shutdown window value.
54 # "Node stale time" affects two related behaviors.
55 # 1. If a compute node has been running for at least this long, but it
56 # isn't paired with an Arvados node, do not shut it down, but leave it alone.
57 # This prevents the node manager from shutting down a node that might
58 # actually be doing work, but is having temporary trouble contacting the
60 # 2. When the Node Manager starts a new compute node, it will try to reuse
61 # an Arvados node that hasn't been updated for this long.
62 node_stale_after = 14400
64 # Scaling factor to be applied to nodes' available RAM size. Usually there's a
65 # variable discrepancy between the advertised RAM value on cloud nodes and the
66 # actual amount available.
67 # If not set, this value will be set to 0.95
68 node_mem_scaling = 0.95
70 # File path for Certificate Authorities
71 certs_file = /etc/ssl/certs/ca-certificates.crt
75 #file = node-manager.log
77 # Log level for most Node Manager messages.
78 # Choose one of DEBUG, INFO, WARNING, ERROR, or CRITICAL.
79 # WARNING lets you know when polling a service fails.
80 # INFO additionally lets you know when a compute node is started or stopped.
83 # You can also set different log levels for specific libraries.
84 # Pykka is the Node Manager's actor library.
85 # Setting this to DEBUG will display tracebacks for uncaught
86 # exceptions in the actors, but it's also very chatty.
89 # Setting apiclient to INFO will log the URL of every Arvados API request.
98 # Accept an untrusted SSL certificate from the API server?
103 driver_class = {driver_class}
105 # Shutdown windows define periods of time when a node may and may not be shut
106 # down. These are windows in full minutes, separated by commas. Counting from
107 # the time the node is booted, the node WILL NOT shut down for N1 minutes; then
108 # it MAY shut down for N2 minutes; then it WILL NOT shut down for N3 minutes;
109 # and so on. For example, "20, 999999" means the node may shut down between
110 # the 20th and 999999th minutes of uptime.
111 # Azure bills by the minute, so it makes sense to agressively shut down idle
112 # nodes. Specify at least two windows. You can add as many as you need beyond
114 shutdown_windows = 1, 999999
117 key = 00000000-0000-0000-0000-000000000000
126 image = fake_image_id
128 # Path to a local ssh key file that will be used to provision new nodes.
131 # the API server to ping
134 # You can define any number of Size sections to list Azure sizes you're willing
135 # to use. The Node Manager should boot the cheapest size(s) that can run jobs
136 # in the queue. You must also provide price per hour as the Azure driver
137 # compute currently does not report prices.
139 # See https://azure.microsoft.com/en-us/pricing/details/virtual-machines/
140 # for a list of known machine types that may be used as a Size parameter.
142 # Each size section MUST define the number of cores are available in this
143 # size class (since libcloud does not provide any consistent API for exposing
145 # You may also want to define the amount of scratch space (expressed
146 # in GB) for Crunch jobs. You can also override Microsoft's provided
147 # data fields by setting them here.