The bleeding edge branch is now called 'main'.
[arvados.git] / doc / install / crunch2-cloud / install-dispatch-cloud.html.textile.liquid
1 ---
2 layout: default
3 navsection: installguide
4 title: Install the cloud dispatcher
5 ...
6 {% comment %}
7 Copyright (C) The Arvados Authors. All rights reserved.
8
9 SPDX-License-Identifier: CC-BY-SA-3.0
10 {% endcomment %}
11
12 {% include 'notebox_begin_warning' %}
13 arvados-dispatch-cloud is only relevant for cloud installations. Skip this section if you are installing an on premises cluster that will spool jobs to Slurm.
14 {% include 'notebox_end' %}
15
16 # "Introduction":#introduction
17 # "Create compute node VM image":#create-image
18 # "Update config.yml":#update-config
19 # "Install arvados-dispatch-cloud":#install-packages
20 # "Start the service":#start-service
21 # "Restart the API server and controller":#restart-api
22 # "Confirm working installation":#confirm-working
23
24 h2(#introduction). Introduction
25
26 The cloud dispatch service is for running containers on cloud VMs. It works with Microsoft Azure and Amazon EC2; future versions will also support Google Compute Engine.
27
28 The cloud dispatch service can run on any node that can connect to the Arvados API service, the cloud provider's API, and the SSH service on cloud VMs.  It is not resource-intensive, so you can run it on the API server node.
29
30 More detail about the internal operation of the dispatcher can be found in the "architecture section":{{site.baseurl}}/architecture/dispatchcloud.html.
31
32 h2(#update-config). Update config.yml
33
34 h3. Configure CloudVMs
35
36 Add or update the following portions of your cluster configuration file, @config.yml@. Refer to "config.defaults.yml":{{site.baseurl}}/admin/config.html for information about additional configuration options. The @DispatchPrivateKey@ should be the *private* key generated in "the previous section":install-compute-node.html#sshkeypair.
37
38 <notextile>
39 <pre><code>    Services:
40       DispatchCloud:
41         InternalURLs:
42           "http://localhost:9006": {}
43     Containers:
44       CloudVMs:
45         # BootProbeCommand is a shell command that succeeds when an instance is ready for service
46         BootProbeCommand: "sudo systemctl status docker"
47
48         <b># --- driver-specific configuration goes here --- see Amazon and Azure examples below ---</b>
49
50       DispatchPrivateKey: |
51         -----BEGIN RSA PRIVATE KEY-----
52         MIIEpQIBAAKCAQEAqXoCzcOBkFQ7w4dvXf9B++1ctgZRqEbgRYL3SstuMV4oawks
53         ttUuxJycDdsPmeYcHsKo8vsEZpN6iYsX6ZZzhkO5nEayUTU8sBjmg1ZCTo4QqKXr
54         FJ+amZ7oYMDof6QEdwl6KNDfIddL+NfBCLQTVInOAaNss7GRrxLTuTV7HcRaIUUI
55         jYg0Ibg8ZZTzQxCvFXXnjseTgmOcTv7CuuGdt91OVdoq8czG/w8TwOhymEb7mQlt
56         lXuucwQvYgfoUgcnTgpJr7j+hafp75g2wlPozp8gJ6WQ2yBWcfqL2aw7m7Ll88Nd
57         [...]
58         oFyAjVoexx0RBcH6BveTfQtJKbktP1qBO4mXo2dP0cacuZEtlAqW9Eb06Pvaw/D9
59         foktmqOY8MyctzFgXBpGTxPliGjqo8OkrOyQP2g+FL7v+Km31Xs61P8=
60         -----END RSA PRIVATE KEY-----
61     InstanceTypes:
62       x1md:
63         ProviderType: x1.medium
64         VCPUs: 8
65         RAM: 64GiB
66         IncludedScratch: 64GB
67         Price: 0.62
68       x1lg:
69         ProviderType: x1.large
70         VCPUs: 16
71         RAM: 128GiB
72         IncludedScratch: 128GB
73         Price: 1.23
74 </code></pre>
75 </notextile>
76
77 h4. Minimal configuration example for Amazon EC2
78
79 The <span class="userinput">ImageID</span> value is the compute node image that was built in "the previous section":install-compute-node.html#aws.
80
81 <notextile>
82 <pre><code>    Containers:
83       CloudVMs:
84         ImageID: <span class="userinput">ami-01234567890abcdef</span>
85         Driver: ec2
86         DriverParameters:
87           # If you are not using an IAM role for authentication, specify access
88           # credentials here. Otherwise, omit or set AccessKeyID and
89           # SecretAccessKey to an empty value.
90           AccessKeyID: XXXXXXXXXXXXXXXXXXXX
91           SecretAccessKey: YYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYY
92
93           SecurityGroupIDs:
94           - sg-0123abcd
95           SubnetID: subnet-0123abcd
96           Region: us-east-1
97           EBSVolumeType: gp2
98           AdminUsername: arvados
99 </code></pre>
100 </notextile>
101
102 h4. Minimal configuration example for Azure
103
104 Using managed disks:
105
106 The <span class="userinput">ImageID</span> value is the compute node image that was built in "the previous section":install-compute-node.html#azure.
107
108 <notextile>
109 <pre><code>    Containers:
110       CloudVMs:
111         ImageID: <span class="userinput">"zzzzz-compute-v1597349873"</span>
112         Driver: azure
113         # (azure) managed disks: set MaxConcurrentInstanceCreateOps to 20 to avoid timeouts, cf
114         # https://docs.microsoft.com/en-us/azure/virtual-machines/linux/capture-image
115         MaxConcurrentInstanceCreateOps: 20
116         DriverParameters:
117           # Credentials.
118           SubscriptionID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
119           ClientID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
120           ClientSecret: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
121           TenantID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
122
123           # Data center where VMs will be allocated
124           Location: centralus
125
126           # The resource group where the VM and virtual NIC will be
127           # created.
128           ResourceGroup: zzzzz
129           NetworkResourceGroup: yyyyy   # only if different from ResourceGroup
130           Network: xxxxx
131           Subnet: xxxxx-subnet-private
132
133           # The resource group where the disk image is stored, only needs to
134           # be specified if it is different from ResourceGroup
135           ImageResourceGroup: aaaaa
136
137 </code></pre>
138 </notextile>
139
140 Azure recommends using managed images. If you plan to start more than 20 VMs simultaneously, Azure recommends using a shared image gallery instead to avoid slowdowns and timeouts during the creation of the VMs.
141
142 Using an image from a shared image gallery:
143
144 <notextile>
145 <pre><code>    Containers:
146       CloudVMs:
147         ImageID: <span class="userinput">"shared_image_gallery_image_definition_name"</span>
148         Driver: azure
149         DriverParameters:
150           # Credentials.
151           SubscriptionID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
152           ClientID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
153           ClientSecret: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
154           TenantID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
155
156           # Data center where VMs will be allocated
157           Location: centralus
158
159           # The resource group where the VM and virtual NIC will be
160           # created.
161           ResourceGroup: zzzzz
162           NetworkResourceGroup: yyyyy   # only if different from ResourceGroup
163           Network: xxxxx
164           Subnet: xxxxx-subnet-private
165
166           # The resource group where the disk image is stored, only needs to
167           # be specified if it is different from ResourceGroup
168           ImageResourceGroup: aaaaa
169
170           # (azure) shared image gallery: the name of the gallery
171           SharedImageGalleryName: "shared_image_gallery_1"
172           # (azure) shared image gallery: the version of the image definition
173           SharedImageGalleryImageVersion: "0.0.1"
174
175 </code></pre>
176 </notextile>
177
178 Using unmanaged disks (deprecated):
179
180 The <span class="userinput">ImageID</span> value is the compute node image that was built in "the previous section":install-compute-node.html#azure.
181
182 <notextile>
183 <pre><code>    Containers:
184       CloudVMs:
185         ImageID: <span class="userinput">"https://zzzzzzzz.blob.core.windows.net/system/Microsoft.Compute/Images/images/zzzzz-compute-osDisk.55555555-5555-5555-5555-555555555555.vhd"</span>
186         Driver: azure
187         DriverParameters:
188           # Credentials.
189           SubscriptionID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
190           ClientID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
191           ClientSecret: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
192           TenantID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
193
194           # Data center where VMs will be allocated
195           Location: centralus
196
197           # The resource group where the VM and virtual NIC will be
198           # created.
199           ResourceGroup: zzzzz
200           NetworkResourceGroup: yyyyy   # only if different from ResourceGroup
201           Network: xxxxx
202           Subnet: xxxxx-subnet-private
203
204           # Where to store the VM VHD blobs
205           StorageAccount: example
206           BlobContainer: vhds
207
208 </code></pre>
209 </notextile>
210
211 Get the @SubscriptionID@ and @TenantID@:
212
213 <pre>
214 $ az account list
215 [
216   {
217     "cloudName": "AzureCloud",
218     "id": "XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXX",
219     "isDefault": true,
220     "name": "Your Subscription",
221     "state": "Enabled",
222     "tenantId": "YYYYYYYY-YYYY-YYYY-YYYYYYYY",
223     "user": {
224       "name": "you@example.com",
225       "type": "user"
226     }
227   }
228 ]
229 </pre>
230
231 You will need to create a "service principal" to use as a delegated authority for API access.
232
233 <notextile><pre><code>$ az ad app create --display-name "Arvados Dispatch Cloud (<span class="userinput">ClusterID</span>)" --homepage "https://arvados.org" --identifier-uris "https://<span class="userinput">ClusterID.example.com</span>" --end-date 2299-12-31 --password <span class="userinput">Your_Password</span>
234 $ az ad sp create "<span class="userinput">appId</span>"
235 (appId is part of the response of the previous command)
236 $ az role assignment create --assignee "<span class="userinput">objectId</span>" --role Owner --scope /subscriptions/{subscriptionId}/
237 (objectId is part of the response of the previous command)
238 </code></pre></notextile>
239
240 Now update your @config.yml@ file:
241
242 @ClientID@ is the 'appId' value.
243
244 @ClientSecret@ is what was provided as <span class="userinput">Your_Password</span>.
245
246 h3. Test your configuration
247
248 Run the @cloudtest@ tool to verify that your configuration works. This creates a new cloud VM, confirms that it boots correctly and accepts your configured SSH private key, and shuts it down.
249
250 <notextile>
251 <pre><code>~$ <span class="userinput">arvados-server cloudtest && echo "OK!"</span>
252 </code></pre>
253 </notextile>
254
255 Refer to the "cloudtest tool documentation":../../admin/cloudtest.html for more information.
256
257 {% assign arvados_component = 'arvados-dispatch-cloud' %}
258
259 {% include 'install_packages' %}
260
261 {% include 'start_service' %}
262
263 {% include 'restart_api' %}
264
265 h2(#confirm-working). Confirm working installation
266
267 On the dispatch node, start monitoring the arvados-dispatch-cloud logs:
268
269 <notextile>
270 <pre><code>~$ <span class="userinput">sudo journalctl -o cat -fu arvados-dispatch-cloud.service</span>
271 </code></pre>
272 </notextile>
273
274 "Make sure to install the arvados/jobs image.":../install-jobs-image.html
275
276 Submit a simple container request:
277
278 <notextile>
279 <pre><code>shell:~$ <span class="userinput">arv container_request create --container-request '{
280   "name":            "test",
281   "state":           "Committed",
282   "priority":        1,
283   "container_image": "arvados/jobs:latest",
284   "command":         ["echo", "Hello, Crunch!"],
285   "output_path":     "/out",
286   "mounts": {
287     "/out": {
288       "kind":        "tmp",
289       "capacity":    1000
290     }
291   },
292   "runtime_constraints": {
293     "vcpus": 1,
294     "ram": 1048576
295   }
296 }'</span>
297 </code></pre>
298 </notextile>
299
300 This command should return a record with a @container_uuid@ field.  Once @arvados-dispatch-cloud@ polls the API server for new containers to run, you should see it dispatch that same container.
301
302 The @arvados-dispatch-cloud@ API provides a list of queued and running jobs and cloud instances. Use your @ManagementToken@ to test the dispatcher's endpoint. For example, when one container is running:
303
304 <notextile>
305 <pre><code>~$ <span class="userinput">curl -sH "Authorization: Bearer $token" http://localhost:9006/arvados/v1/dispatch/containers</span>
306 {
307   "items": [
308     {
309       "container": {
310         "uuid": "zzzzz-dz642-hdp2vpu9nq14tx0",
311         ...
312         "state": "Running",
313         "scheduling_parameters": {
314           "partitions": null,
315           "preemptible": false,
316           "max_run_time": 0
317         },
318         "exit_code": 0,
319         "runtime_status": null,
320         "started_at": null,
321         "finished_at": null
322       },
323       "instance_type": {
324         "Name": "Standard_D2s_v3",
325         "ProviderType": "Standard_D2s_v3",
326         "VCPUs": 2,
327         "RAM": 8589934592,
328         "Scratch": 16000000000,
329         "IncludedScratch": 16000000000,
330         "AddedScratch": 0,
331         "Price": 0.11,
332         "Preemptible": false
333       }
334     }
335   ]
336 }
337 </code></pre>
338 </notextile>
339
340 A similar request can be made to the @http://localhost:9006/arvados/v1/dispatch/instances@ endpoint.
341
342 When the container finishes, the dispatcher will log it.
343
344 After the container finishes, you can get the container record by UUID *from a shell server* to see its results:
345
346 <notextile>
347 <pre><code>shell:~$ <span class="userinput">arv get <b>zzzzz-dz642-hdp2vpu9nq14tx0</b></span>
348 {
349  ...
350  "exit_code":0,
351  "log":"a01df2f7e5bc1c2ad59c60a837e90dc6+166",
352  "output":"d41d8cd98f00b204e9800998ecf8427e+0",
353  "state":"Complete",
354  ...
355 }
356 </code></pre>
357 </notextile>
358
359 You can use standard Keep tools to view the container's output and logs from their corresponding fields.  For example, to see the logs from the collection referenced in the @log@ field:
360
361 <notextile>
362 <pre><code>~$ <span class="userinput">arv keep ls <b>a01df2f7e5bc1c2ad59c60a837e90dc6+166</b></span>
363 ./crunch-run.txt
364 ./stderr.txt
365 ./stdout.txt
366 ~$ <span class="userinput">arv-get <b>a01df2f7e5bc1c2ad59c60a837e90dc6+166</b>/stdout.txt</span>
367 2016-08-05T13:53:06.201011Z Hello, Crunch!
368 </code></pre>
369 </notextile>
370
371 If the container does not dispatch successfully, refer to the @arvados-dispatch-cloud@ logs for information about why it failed.