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 or LSF.
  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 Example policy for the IAM role used by the cloud dispatcher:
 103
 104 <notextile>
 105 <pre>
 106 {
 107     "Version": "2012-10-17",
 108     "Id": "arvados-dispatch-cloud policy",
 109     "Statement": [
 110         {
 111             "Effect": "Allow",
 112             "Action": [
 113                 "iam:PassRole",
 114                 "ec2:DescribeKeyPairs",
 115                 "ec2:ImportKeyPair",
 116                 "ec2:RunInstances",
 117                 "ec2:DescribeInstances",
 118                 "ec2:CreateTags",
 119                 "ec2:TerminateInstances"
 120             ],
 121             "Resource": "*"
 122         }
 123     ]
 124 }
 125 </pre>
 126 </notextile>
 127
 128 h4. Minimal configuration example for Azure
 129
 130 Using managed disks:
 131
 132 The <span class="userinput">ImageID</span> value is the compute node image that was built in "the previous section":install-compute-node.html#azure.
 133
 134 <notextile>
 135 <pre><code>    Containers:
 136       CloudVMs:
 137         ImageID: <span class="userinput">"zzzzz-compute-v1597349873"</span>
 138         Driver: azure
 139         # (azure) managed disks: set MaxConcurrentInstanceCreateOps to 20 to avoid timeouts, cf
 140         # https://docs.microsoft.com/en-us/azure/virtual-machines/linux/capture-image
 141         MaxConcurrentInstanceCreateOps: 20
 142         DriverParameters:
 143           # Credentials.
 144           SubscriptionID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 145           ClientID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 146           ClientSecret: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 147           TenantID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 148
 149           # Data center where VMs will be allocated
 150           Location: centralus
 151
 152           # The resource group where the VM and virtual NIC will be
 153           # created.
 154           ResourceGroup: zzzzz
 155           NetworkResourceGroup: yyyyy   # only if different from ResourceGroup
 156           Network: xxxxx
 157           Subnet: xxxxx-subnet-private
 158
 159           # The resource group where the disk image is stored, only needs to
 160           # be specified if it is different from ResourceGroup
 161           ImageResourceGroup: aaaaa
 162
 163 </code></pre>
 164 </notextile>
 165
 166 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.
 167
 168 Using an image from a shared image gallery:
 169
 170 <notextile>
 171 <pre><code>    Containers:
 172       CloudVMs:
 173         ImageID: <span class="userinput">"shared_image_gallery_image_definition_name"</span>
 174         Driver: azure
 175         DriverParameters:
 176           # Credentials.
 177           SubscriptionID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 178           ClientID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 179           ClientSecret: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 180           TenantID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 181
 182           # Data center where VMs will be allocated
 183           Location: centralus
 184
 185           # The resource group where the VM and virtual NIC will be
 186           # created.
 187           ResourceGroup: zzzzz
 188           NetworkResourceGroup: yyyyy   # only if different from ResourceGroup
 189           Network: xxxxx
 190           Subnet: xxxxx-subnet-private
 191
 192           # The resource group where the disk image is stored, only needs to
 193           # be specified if it is different from ResourceGroup
 194           ImageResourceGroup: aaaaa
 195
 196           # (azure) shared image gallery: the name of the gallery
 197           SharedImageGalleryName: "shared_image_gallery_1"
 198           # (azure) shared image gallery: the version of the image definition
 199           SharedImageGalleryImageVersion: "0.0.1"
 200
 201 </code></pre>
 202 </notextile>
 203
 204 Using unmanaged disks (deprecated):
 205
 206 The <span class="userinput">ImageID</span> value is the compute node image that was built in "the previous section":install-compute-node.html#azure.
 207
 208 <notextile>
 209 <pre><code>    Containers:
 210       CloudVMs:
 211         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>
 212         Driver: azure
 213         DriverParameters:
 214           # Credentials.
 215           SubscriptionID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 216           ClientID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 217           ClientSecret: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 218           TenantID: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
 219
 220           # Data center where VMs will be allocated
 221           Location: centralus
 222
 223           # The resource group where the VM and virtual NIC will be
 224           # created.
 225           ResourceGroup: zzzzz
 226           NetworkResourceGroup: yyyyy   # only if different from ResourceGroup
 227           Network: xxxxx
 228           Subnet: xxxxx-subnet-private
 229
 230           # Where to store the VM VHD blobs
 231           StorageAccount: example
 232           BlobContainer: vhds
 233
 234 </code></pre>
 235 </notextile>
 236
 237 Get the @SubscriptionID@ and @TenantID@:
 238
 239 <pre>
 240 $ az account list
 241 [
 242   {
 243     "cloudName": "AzureCloud",
 244     "id": "XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXX",
 245     "isDefault": true,
 246     "name": "Your Subscription",
 247     "state": "Enabled",
 248     "tenantId": "YYYYYYYY-YYYY-YYYY-YYYYYYYY",
 249     "user": {
 250       "name": "you@example.com",
 251       "type": "user"
 252     }
 253   }
 254 ]
 255 </pre>
 256
 257 You will need to create a "service principal" to use as a delegated authority for API access.
 258
 259 <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>
 260 $ az ad sp create "<span class="userinput">appId</span>"
 261 (appId is part of the response of the previous command)
 262 $ az role assignment create --assignee "<span class="userinput">objectId</span>" --role Owner --scope /subscriptions/{subscriptionId}/
 263 (objectId is part of the response of the previous command)
 264 </code></pre></notextile>
 265
 266 Now update your @config.yml@ file:
 267
 268 @ClientID@ is the 'appId' value.
 269
 270 @ClientSecret@ is what was provided as <span class="userinput">Your_Password</span>.
 271
 272 h3. Test your configuration
 273
 274 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.
 275
 276 <notextile>
 277 <pre><code>~$ <span class="userinput">arvados-server cloudtest && echo "OK!"</span>
 278 </code></pre>
 279 </notextile>
 280
 281 Refer to the "cloudtest tool documentation":../../admin/cloudtest.html for more information.
 282
 283 {% assign arvados_component = 'arvados-dispatch-cloud' %}
 284
 285 {% include 'install_packages' %}
 286
 287 {% include 'start_service' %}
 288
 289 {% include 'restart_api' %}
 290
 291 h2(#confirm-working). Confirm working installation
 292
 293 On the dispatch node, start monitoring the arvados-dispatch-cloud logs:
 294
 295 <notextile>
 296 <pre><code>~$ <span class="userinput">sudo journalctl -o cat -fu arvados-dispatch-cloud.service</span>
 297 </code></pre>
 298 </notextile>
 299
 300 "Make sure to install the arvados/jobs image.":../install-jobs-image.html
 301
 302 Submit a simple container request:
 303
 304 <notextile>
 305 <pre><code>shell:~$ <span class="userinput">arv container_request create --container-request '{
 306   "name":            "test",
 307   "state":           "Committed",
 308   "priority":        1,
 309   "container_image": "arvados/jobs:latest",
 310   "command":         ["echo", "Hello, Crunch!"],
 311   "output_path":     "/out",
 312   "mounts": {
 313     "/out": {
 314       "kind":        "tmp",
 315       "capacity":    1000
 316     }
 317   },
 318   "runtime_constraints": {
 319     "vcpus": 1,
 320     "ram": 1048576
 321   }
 322 }'</span>
 323 </code></pre>
 324 </notextile>
 325
 326 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.
 327
 328 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:
 329
 330 <notextile>
 331 <pre><code>~$ <span class="userinput">curl -sH "Authorization: Bearer $token" http://localhost:9006/arvados/v1/dispatch/containers</span>
 332 {
 333   "items": [
 334     {
 335       "container": {
 336         "uuid": "zzzzz-dz642-hdp2vpu9nq14tx0",
 337         ...
 338         "state": "Running",
 339         "scheduling_parameters": {
 340           "partitions": null,
 341           "preemptible": false,
 342           "max_run_time": 0
 343         },
 344         "exit_code": 0,
 345         "runtime_status": null,
 346         "started_at": null,
 347         "finished_at": null
 348       },
 349       "instance_type": {
 350         "Name": "Standard_D2s_v3",
 351         "ProviderType": "Standard_D2s_v3",
 352         "VCPUs": 2,
 353         "RAM": 8589934592,
 354         "Scratch": 16000000000,
 355         "IncludedScratch": 16000000000,
 356         "AddedScratch": 0,
 357         "Price": 0.11,
 358         "Preemptible": false
 359       }
 360     }
 361   ]
 362 }
 363 </code></pre>
 364 </notextile>
 365
 366 A similar request can be made to the @http://localhost:9006/arvados/v1/dispatch/instances@ endpoint.
 367
 368 When the container finishes, the dispatcher will log it.
 369
 370 After the container finishes, you can get the container record by UUID *from a shell server* to see its results:
 371
 372 <notextile>
 373 <pre><code>shell:~$ <span class="userinput">arv get <b>zzzzz-dz642-hdp2vpu9nq14tx0</b></span>
 374 {
 375  ...
 376  "exit_code":0,
 377  "log":"a01df2f7e5bc1c2ad59c60a837e90dc6+166",
 378  "output":"d41d8cd98f00b204e9800998ecf8427e+0",
 379  "state":"Complete",
 380  ...
 381 }
 382 </code></pre>
 383 </notextile>
 384
 385 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:
 386
 387 <notextile>
 388 <pre><code>~$ <span class="userinput">arv keep ls <b>a01df2f7e5bc1c2ad59c60a837e90dc6+166</b></span>
 389 ./crunch-run.txt
 390 ./stderr.txt
 391 ./stdout.txt
 392 ~$ <span class="userinput">arv-get <b>a01df2f7e5bc1c2ad59c60a837e90dc6+166</b>/stdout.txt</span>
 393 2016-08-05T13:53:06.201011Z Hello, Crunch!
 394 </code></pre>
 395 </notextile>
 396
 397 If the container does not dispatch successfully, refer to the @arvados-dispatch-cloud@ logs for information about why it failed.