1 # Copyright (C) The Arvados Authors. All rights reserved.
3 # SPDX-License-Identifier: AGPL-3.0
5 # Azure configuration for Arvados Node Manager.
6 # All times are in seconds unless specified otherwise.
9 # The management server responds to http://addr:port/status.json with
10 # a snapshot of internal state.
12 # Management server listening address (default 127.0.0.1)
15 # Management server port number (default -1, server is disabled)
19 # The dispatcher can customize the start and stop procedure for
20 # cloud nodes. For example, the SLURM dispatcher drains nodes
21 # through SLURM before shutting them down.
24 # Node Manager will ensure that there are at least this many nodes running at
25 # all times. If node manager needs to start new idle nodes for the purpose of
26 # satisfying min_nodes, it will use the cheapest node type. However, depending
27 # on usage patterns, it may also satisfy min_nodes by keeping alive some
28 # more-expensive nodes
31 # Node Manager will not start any compute nodes when at least this
35 # Upper limit on rate of spending (in $/hr), will not boot additional nodes
36 # if total price of already running nodes meets or exceeds this threshold.
37 # default 0 means no limit.
40 # Poll Azure nodes and Arvados for new information every N seconds.
43 # Polls have exponential backoff when services fail to respond.
44 # This is the longest time to wait between polls.
47 # If Node Manager can't succesfully poll a service for this long,
48 # it will never start or stop compute nodes, on the assumption that its
49 # information is too outdated.
50 poll_stale_after = 600
52 # If Node Manager boots a cloud node, and it does not pair with an Arvados
53 # node before this long, assume that there was a cloud bootstrap failure and
54 # shut it down. Note that normal shutdown windows apply (see the Cloud
55 # section), so this should be shorter than the first shutdown window value.
56 boot_fail_after = 1800
58 # "Node stale time" affects two related behaviors.
59 # 1. If a compute node has been running for at least this long, but it
60 # isn't paired with an Arvados node, do not shut it down, but leave it alone.
61 # This prevents the node manager from shutting down a node that might
62 # actually be doing work, but is having temporary trouble contacting the
64 # 2. When the Node Manager starts a new compute node, it will try to reuse
65 # an Arvados node that hasn't been updated for this long.
66 node_stale_after = 14400
68 # Scaling factor to be applied to nodes' available RAM size. Usually there's a
69 # variable discrepancy between the advertised RAM value on cloud nodes and the
70 # actual amount available.
71 # If not set, this value will be set to 0.95
72 node_mem_scaling = 0.95
74 # File path for Certificate Authorities
75 certs_file = /etc/ssl/certs/ca-certificates.crt
79 file = /var/log/arvados/node-manager.log
81 # Log level for most Node Manager messages.
82 # Choose one of DEBUG, INFO, WARNING, ERROR, or CRITICAL.
83 # WARNING lets you know when polling a service fails.
84 # INFO additionally lets you know when a compute node is started or stopped.
87 # You can also set different log levels for specific libraries.
88 # Pykka is the Node Manager's actor library.
89 # Setting this to DEBUG will display tracebacks for uncaught
90 # exceptions in the actors, but it's also very chatty.
93 # Setting apiclient to INFO will log the URL of every Arvados API request.
97 host = zyxwv.arvadosapi.com
100 jobs_queue = yes # Get work request from Arvados jobs queue (jobs API)
101 slurm_queue = yes # Get work request from squeue (containers API)
103 # Accept an untrusted SSL certificate from the API server?
109 # Shutdown windows define periods of time when a node may and may not be shut
110 # down. These are windows in full minutes, separated by commas. Counting from
111 # the time the node is booted, the node WILL NOT shut down for N1 minutes; then
112 # it MAY shut down for N2 minutes; then it WILL NOT shut down for N3 minutes;
113 # and so on. For example, "20, 999999" means the node may shut down between
114 # the 20th and 999999th minutes of uptime.
115 # Azure bills by the minute, so it makes sense to agressively shut down idle
116 # nodes. Specify at least two windows. You can add as many as you need beyond
118 shutdown_windows = 20, 999999
121 # Use "azure account list" with the azure CLI to get these values.
122 tenant_id = 00000000-0000-0000-0000-000000000000
123 subscription_id = 00000000-0000-0000-0000-000000000000
125 # The following directions are based on
126 # https://azure.microsoft.com/en-us/documentation/articles/resource-group-authenticate-service-principal/
128 # azure config mode arm
129 # azure ad app create --name "<Your Application Display Name>" --home-page "<https://YourApplicationHomePage>" --identifier-uris "<https://YouApplicationUri>" --password <Your_Password>
130 # azure ad sp create "<Application_Id>"
131 # azure role assignment create --objectId "<Object_Id>" -o Owner -c /subscriptions/{subscriptionId}/
133 # Use <Application_Id> for "key" and the <Your_Password> for "secret"
135 key = 00000000-0000-0000-0000-000000000000
141 # The resource group in which the compute node virtual machines will be created
143 ex_resource_group = ArvadosResourceGroup
146 # The image id, in the form "Publisher:Offer:SKU:Version"
147 image = Canonical:UbuntuServer:14.04.3-LTS:14.04.201508050
149 # Path to a local ssh key file that will be used to provision new nodes.
150 ssh_key = /home/arvadosuser/.ssh/id_rsa.pub
152 # The account name for the admin user that will be provisioned on new nodes.
153 ex_user_name = arvadosuser
155 # The Azure storage account that will be used to store the node OS disk images.
156 ex_storage_account = arvadosstorage
158 # The virtual network the VMs will be associated with.
159 ex_network = ArvadosNetwork
161 # Optional subnet of the virtual network.
165 tag_arvados-class = dynamic-compute
168 # the API server to ping
169 ping_host = hostname:port
171 # You can define any number of Size sections to list Azure sizes you're willing
172 # to use. The Node Manager should boot the cheapest size(s) that can run jobs
173 # in the queue. You must also provide price per hour as the Azure driver
174 # compute currently does not report prices.
176 # See https://azure.microsoft.com/en-us/pricing/details/virtual-machines/
177 # for a list of known machine types that may be used as a Size parameter.
179 # Each size section MUST define the number of cores are available in this
180 # size class (since libcloud does not provide any consistent API for exposing
182 # You may also want to define the amount of scratch space (expressed
183 # in GB) for Crunch jobs. You can also override Microsoft's provided
184 # data fields by setting them here.