Project

General

Profile

Dispatching containers to cloud VMs » History » Version 57

Tom Clegg, 02/20/2019 09:31 PM

1 1 Tom Clegg
h1. Dispatching containers to cloud VMs
2
3 6 Tom Clegg
(Draft)
4 1 Tom Clegg
5 11 Tom Clegg
{{>toc}}
6
7 7 Tom Clegg
h2. Component name / purpose
8 1 Tom Clegg
9 53 Tom Clegg
arvados-dispatch-cloud runs Arvados user containers on generic public cloud infrastructure by automatically creating and destroying VMs of various sizes according to demand, preparing the VMs' runtime environments, and running containers on them.
10 1 Tom Clegg
11 49 Tom Clegg
h2. Overview of operation
12
13
The dispatcher waits for containers to appear in the queue, and runs them on appropriately sized cloud VMs. When there are no idle cloud VMs with the desired size, the dispatcher brings up more VMs using the cloud provider's API. The dispatcher also shuts down idle VMs that exceed the configured idle timer -- and sooner if the provider starts refusing to create new VMs.
14
15
h2. Interaction with other components
16
17
Controller (backed by RailsAPI and PostgreSQL) supplies the container queue: which containers the system should be trying to execute (or cancel) at any given time.
18
19
The cloud provider's API supplies a list of VMs that exist (or are being created) at a given time and their network addresses, accepts orders to create new VMs, updates instance tags, and (optionally, depending on the driver) obtains the VMs' SSH server public keys.
20
21 57 Tom Clegg
The SSH server on each cloud VM allows the dispatcher to authenticate with a private key and execute shell commands as root (either directly or via sudo).
22 49 Tom Clegg
23
h2. Instance tags
24
25
The dispatcher relies on the cloud provider's tagging feature to persist state across server restarts.
26
* {"InstanceType": "foo"} indicates that the instance was created with the specs from the instance type named "foo" in the cluster configuration file.
27 50 Tom Clegg
* {"IdleBehavior": "hold"} indicates that the management API has been used to put the instance in "hold" state.
28 49 Tom Clegg
29
Provider-specific drivers (Amazon, Google, Azure) determine exactly how these tags are encoded in the cloud API, and can use tags to persist their own internal state as well. For example, a driver might save tags named "Arvados-DispatchCloud-InstanceType" rather than just "InstanceType".
30
31 23 Tom Clegg
h2. Deployment
32 24 Tom Clegg
33 53 Tom Clegg
*Where to install:* The arvados-dispatch-cloud process can run anywhere, as long as it has network access to the Arvados controller, the cloud provider's API, and the worker VMs. Each Arvados cluster should run only one arvados-dispatch-cloud process.
34 9 Tom Clegg
* Future versions will support multiple dispatchers.
35 1 Tom Clegg
36 9 Tom Clegg
*Dispatcher's SSH key:* The operator must generate an SSH key pair for the dispatcher to use when connecting to cloud VMs. The private key is stored (without a passphrase) in the cluster configuration file. It does not need to be saved in @~/.ssh/@.
37 1 Tom Clegg
38 56 Tom Clegg
*Cloud VM image:* The operator must provide a VM image with an SSH server on a port reachable by the dispatcher (default 22, configurable per cluster). The dispatcher's SSH public key must be listed in @/root/.ssh/authorized_keys@. The image should also include systemd-cat (part of systemd) and suitable versions of docker and crunch-run. The @/var/lock@ directory must be available for lockfiles with names matching "@crunch-run-*.*@".
39 1 Tom Clegg
* It is possible to install docker and crunch-run using a custom boot probe command, but pre-installing is more efficient.
40
* Future versions will automatically sync the crunch-run binary from the dispatcher host to each worker node.
41 56 Tom Clegg
* The Azure driver creates a new admin user account and installs the SSH public key by itself so @/root/.ssh/authorized_keys@ is not needed. The VM image must include @sudo@.
42 1 Tom Clegg
43 6 Tom Clegg
*Cloud provider account:* The dispatcher uses cloud provider credentials to create and delete VMs and other cloud resources. An Arvados user can create an arbitrary number of long-running containers, and the dispatcher will try to run all of them. Currently the dispatcher does not enforce any resource limits of its own, so the operator must ensure the cloud provider itself is enforcing a suitable quota.
44 53 Tom Clegg
45 52 Tom Clegg
*Migrating from nodemanager/SLURM:* When VM images, SSH keys, and configuration files are ready, disable nodemanager and crunch-dispatch-slurm. Install arvados-dispatch-cloud deb/rpm package. Confirm success with @systemctl status arvados-dispatch-cloud@ and @journalctl -fu arvados-dispatch-cloud@. See [[Migrating from arvados-node-manager to arvados-dispatch-cloud]].
46 1 Tom Clegg
47 6 Tom Clegg
h2. Configuration
48 1 Tom Clegg
49 42 Tom Clegg
Arvados [[Cluster configuration]] (currently a file in /etc) supplies cloud provider credentials, allowed node types, spending limits/policies, etc.
50 1 Tom Clegg
51 6 Tom Clegg
<pre><code class="yaml">
52
    CloudVMs:
53 1 Tom Clegg
      BootProbeCommand: "docker ps -q"
54 42 Tom Clegg
      SSHPort: 22
55 27 Tom Clegg
      SyncInterval: 1m    # how often to get list of active instances from cloud provider
56 8 Tom Clegg
      TimeoutIdle: 1m     # shutdown if idle longer than this
57
      TimeoutBooting: 10m # shutdown if exists longer than this without running BootProbeCommand successfully
58
      TimeoutProbe: 2m    # shutdown if (after booting) communication fails longer than this, even if ctrs are running
59
      TimeoutShutdown: 1m # shutdown again if node still exists this long after shutdown
60 6 Tom Clegg
      Driver: Amazon
61 8 Tom Clegg
      DriverParameters:   # following configs are driver dependent
62 6 Tom Clegg
        Region: us-east-1
63
        APITimeout: 20s
64 39 Tom Clegg
        AWSAccessKeyID: abcdef
65
        AWSSecretAccessKey: abcdefghijklmnopqrstuvwxyz
66 6 Tom Clegg
        ImageID: ami-0123456789abcdef0
67 1 Tom Clegg
        SubnetID: subnet-01234567
68
        SecurityGroups: sg-01234567
69 8 Tom Clegg
    Dispatch:
70
      StaleLockTimeout: 1m     # after restart, time to wait for workers to come up before abandoning locks from previous run
71
      PollInterval: 1m         # how often to get latest queue from arvados controller
72
      ProbeInterval: 10s       # how often to probe each instance for current status/vital signs
73
      MaxProbesPerSecond: 1000 # limit total probe rate for dispatch process (across all instances)
74
      PrivateKey: |            # SSH key able to log in as root@ worker VMs
75
        -----BEGIN RSA PRIVATE KEY-----
76
        MIIEowIBAAKCAQEAqYm4XsQHm8sBSZFwUX5VeW1OkGsfoNzcGPG2nzzYRhNhClYZ
77
        0ABHhUk82HkaC/8l6d/jpYTf42HrK42nNQ0r0Yzs7qw8yZMQioK4Yk+kFyVLF78E
78
        GRG4pGAWXFs6pUchs/lm8fo9zcda4R3XeqgI+NO+nEERXmdRJa1FhI+Za3/S/+CV
79
        mg+6O00wZz2+vKmDPptGN4MCKmQOCKsMJts7wSZGyVcTtdNv7jjfr6yPAIOIL8X7
80
        LtarBCFaK/pD7uWll/Uj7h7D8K48nIZUrvBJJjXL8Sm4LxCNoz3Z83k8J5ZzuDRD
81
        gRiQe/C085mhO6VL+2fypDLwcKt1tOL8fI81MwIDAQABAoIBACR3tEnmHsDbNOav
82
        Oxq8cwRQh9K2yDHg8BMJgz/TZa4FIx2HEbxVIw0/iLADtJ+Z/XzGJQCIiWQuvtg6
83
        exoFQESt7JUWRWkSkj9JCQJUoTY9Vl7APtBpqG7rIEQzd3TvzQcagZNRQZQO6rR7
84
        p8sBdBSZ72lK8cJ9tM3G7Kor/VNK7KgRZFNhEWnmvEa3qMd4hzDcQ4faOn7C9NZK
85
        dwJAuJVVfwOLlOORYcyEkvksLaDOK2DsB/p0AaCpfSmThRbBKN5fPXYaKgUdfp3w
86
        70Hpp27WWymb1cgjyqSH3DY+V/kvid+5QxgxCBRq865jPLn3FFT9bWEVS/0wvJRj
87
        iMIRrjECgYEA4Ffv9rBJXqVXonNQbbstd2PaprJDXMUy9/UmfHL6pkq1xdBeuM7v
88
        yf2ocXheA8AahHtIOhtgKqwv/aRhVK0ErYtiSvIk+tXG+dAtj/1ZAKbKiFyxjkZV
89
        X72BH7cTlR6As5SRRfWM/HaBGEgED391gKsI5PyMdqWWdczT5KfxAksCgYEAwXYE
90
        ewPmV1GaR5fbh2RupoPnUJPMj36gJCnwls7sGaXDQIpdlq56zfKgrLocGXGgj+8f
91
        QH7FHTJQO15YCYebtsXWwB3++iG43gVlJlecPAydsap2CCshqNWC5JU5pan0QzsP
92
        exzNzWqfUPSbTkR2SRaN+MenZo2Y/WqScOAth7kCgYBgVoLujW9EXH5QfXJpXLq+
93
        jTvE38I7oVcs0bJwOLPYGzcJtlwmwn6IYAwohgbhV2pLv+EZSs42JPEK278MLKxY
94
        lgVkp60npgunFTWroqDIvdc1TZDVxvA8h9VeODEJlSqxczgbMcIUXBM9yRctTI+5
95
        7DiKlMUA4kTFW2sWwuOlFwKBgGXvrYS0FVbFJKm8lmvMu5D5x5RpjEu/yNnFT4Pn
96
        G/iXoz4Kqi2PWh3STl804UF24cd1k94D7hDoReZCW9kJnz67F+C67XMW+bXi2d1O
97
        JIBvlVfcHb1IHMA9YG7ZQjrMRmx2Xj3ce4RVPgUGHh8ra7gvLjd72/Tpf0doNClN
98
        ti/hAoGBAMW5D3LhU05LXWmOqpeT4VDgqk4MrTBcstVe7KdVjwzHrVHCAmI927vI
99 1 Tom Clegg
        pjpphWzpC9m3x4OsTNf8m+g6H7f3IiQS0aiFNtduXYlcuT5FHS2fSATTzg5PBon9
100
        1E6BudOve+WyFyBs7hFWAqWFBdWujAl4Qk5Ek09U2ilFEPE7RTgJ
101
        -----END RSA PRIVATE KEY-----
102 9 Tom Clegg
    InstanceTypes:
103
    - Name: m4.large
104
      VCPUs: 2
105
      RAM: 7782000000
106
      Scratch: 32000000000
107
      Price: 0.1
108
    - Name: m4.large.spot
109
      Preemptible: true
110
      VCPUs: 2
111
      RAM: 7782000000
112
      Scratch: 32000000000
113
      Price: 0.1
114
    - Name: m4.xlarge
115
      VCPUs: 4
116
      RAM: 15564000000
117
      Scratch: 80000000000
118
      Price: 0.2
119
    - Name: m4.xlarge.spot
120
      Preemptible: true
121
      VCPUs: 4
122
      RAM: 15564000000
123
      Scratch: 80000000000
124
      Price: 0.2
125
    - Name: m4.2xlarge
126
      VCPUs: 8
127
      RAM: 31129000000
128
      Scratch: 160000000000
129
      Price: 0.4
130
    - Name: m4.2xlarge.spot
131
      Preemptible: true
132
      VCPUs: 8
133
      RAM: 31129000000
134
      Scratch: 160000000000
135
      Price: 0.4
136 6 Tom Clegg
</code></pre>
137 1 Tom Clegg
138 10 Tom Clegg
h2. Management API
139 1 Tom Clegg
140 10 Tom Clegg
APIs for monitoring/diagnostics/control are available via HTTP on a configurable address/port. Request headers must include "Authorization: Bearer {management token}".
141
142
Responses are JSON-encoded and resemble other Arvados APIs:
143
<pre><code class="json">
144
{
145 43 Tom Clegg
  "items": [
146 10 Tom Clegg
    {
147 43 Tom Clegg
      "name": "...",
148 10 Tom Clegg
      ...
149
    },
150
    ...
151
  ]
152
}
153
</code></pre>
154
155
@GET /arvados/v1/dispatch/instances@ lists cloud VMs. Each returned item includes:
156
* provider's instance ID
157
* hourly price (from configuration file)
158
* instance type (from configuration file)
159
* instance type (from provider's menu)
160
* UUID of the current / most recent container attempted (if known)
161
* time last container finished (or boot time, if nothing run yet)
162
163
@GET /arvados/v1/dispatch/containers@ lists queued/locked/running containers. Each returned item includes:
164
* container UUID
165
* container state (Queued/Locked/Running/Complete/Cancelled)
166
* desired instance type
167
* time appeared in queue
168
* time started (if started)
169
170 54 Tom Clegg
@POST /arvados/v1/dispatch/instances/hold?instance_id=X@ puts an instance in "hold" state.
171 34 Tom Clegg
* if the instance is currently running a container, it is allowed to continue
172
* no further containers will be scheduled on the instance
173
* the instance will not be shut down automatically
174
175 54 Tom Clegg
@POST /arvados/v1/dispatch/instances/drain?instance_id=X@ puts an instance in "drain" state.
176 1 Tom Clegg
* if the instance is currently running a container, it is allowed to continue
177
* no further containers will be scheduled on the instance
178 34 Tom Clegg
* the instance will be shut down automatically when all containers finish
179 1 Tom Clegg
180 55 Tom Clegg
@POST /arvados/v1/dispatch/instances/run?instance_id=X@ puts an instance in the default "run" state.
181
* if the instance is currently running a container, it is allowed to continue
182
* more containers will be scheduled on the instance when it becomes available
183
* the instance will be shut down automatically when it exceeds the configured idle timeout
184
185 54 Tom Clegg
&dagger;@POST /arvados/v1/dispatch/instances/kill?instance_id=X@ shuts down an instance immediately.
186 1 Tom Clegg
* the instance is terminated immediately via cloud API
187 34 Tom Clegg
* SIGTERM is sent to the container if one is running, but no effort is made to give it time to end gracefully before terminating the instance
188 54 Tom Clegg
189
&dagger;@POST /arvados/v1/dispatch/loglevel?level=debug@ sets the logging threshold to "debug" or "info".
190
* @.../loglevel?level=debug@ enables debug logs
191
* @.../loglevel?level=info@ disables debug logs
192 47 Tom Clegg
193 10 Tom Clegg
h2. Metrics
194 13 Tom Clegg
195 10 Tom Clegg
Metrics are available via HTTP on a configurable address/port (conventionally :9005). Request headers must include "Authorization: Bearer {management token}".
196
197 13 Tom Clegg
Metrics include:
198 1 Tom Clegg
* (gauge) number of existing VMs
199 35 Tom Clegg
* (gauge) total hourly price of all existing VMs
200 46 Tom Clegg
* (gauge) total VCPUs and memory in all existing VMs
201 1 Tom Clegg
* (gauge) total VCPUs and memory allocated to containers
202
* (gauge) number of containers running
203 46 Tom Clegg
* &dagger;(gauge) number of containers allocated to VMs but not started yet (because VMs are pending/booting)
204
* &dagger;(gauge) number of containers not allocated to VMs (because provider quota is reached)
205
* &dagger;(gauge) total hourly price of VMs, partitioned by allocation state (running, booting/idle, adminhold)
206
* &dagger;(summary) time elapsed between VM creation and first successful SSH connection to that VM
207
* &dagger;(summary) time elapsed between first successful SSH connection on a VM and ready to run a container on that VM
208
* &dagger;(summary) time elapsed between first shutdown attempt on a VM and its disappearance from the provider listing
209
210
&dagger; not yet implemented
211
212 14 Tom Clegg
h2. Logs
213 20 Tom Clegg
214 16 Tom Clegg
For purposes of troubleshooting, a JSON-formatted log entry is printed on stderr when...
215 20 Tom Clegg
216 44 Tom Clegg
|                                                              |... if loglevel &ge; ...|...including timestamp and...|
217
|a new instance is created/ordered                             |info                    |instance type name|
218
|an instance appears on the provider's list of instances       |info                    |instance ID|
219
|an instance's boot probe succeeds                             |info                    |instance ID|
220
|an instance is shut down after boot timeout                   |warn                    |instance ID, &dagger;stdout/stderr/error from last boot probe attempt|
221
|an instance shutdown is requested                             |info                    |instance ID|
222
|an instance disappears from the provider's list of instances  |info                    |instance ID and previous state (booting/idle/shutdown)|
223
|a cloud provider API or driver error occurs                   |error                   |provider/driver's error message|
224
|a new container appears in the Arvados queue                  |&dagger;info            |container UUID, desired instance type name|
225 45 Tom Clegg
|a container is locked by the dispatcher                       |debug                   |container UUID|
226 44 Tom Clegg
|a crunch-run process is started on an instance                |info                    |container UUID, instance ID, crunch-run PID|
227
|a crunch-run process fails to start on an instance            |info                    |container UUID, instance ID, stdout/stderr/exitcode|
228
|a crunch-run process ends                                     |info                    |container UUID, instance ID|
229
|an active container's state changes to Complete or Cancelled  |info                    |container UUID, new state|
230
|an active container is requeued after being locked            |info                    |container UUID|  
231
|an Arvados API error occurs                                   |warn                    |error message|
232 16 Tom Clegg
233 44 Tom Clegg
&dagger; not yet implemented
234 1 Tom Clegg
235 51 Tom Clegg
Example log entries from test suite (note test suite uses text formatting, production logging uses JSON formatting):
236
<pre>
237
INFO[0000] creating new instance                         ContainerUUID=zzzzz-dz642-000000000000160 InstanceType=type8
238
INFO[0000] instance appeared in cloud                    IdleBehavior=run Instance=stub-providertype8-6ec34c367674cb74 InstanceType=type8 State=booting
239
INFO[0000] boot probe succeeded                          Command=true Instance=stub-providertype8-6ec34c367674cb74 InstanceType=type8 stderr= stdout=
240
INFO[0000] instance booted; will try probeRunning        Instance=stub-providertype8-6ec34c367674cb74 InstanceType=type8 ProbeStart="2019-02-05 15:49:49.183431341 -0500 EST m=+0.126074285"
241
INFO[0000] probes succeeded, instance is in service      Instance=stub-providertype8-6ec34c367674cb74 InstanceType=type8 ProbeStart="2019-02-05 15:49:49.183431341 -0500 EST m=+0.126074285" RunningContainers=0 State=idle
242
INFO[0000] crunch-run process started                    ContainerUUID=zzzzz-dz642-000000000000160 Instance=stub-providertype8-6ec34c367674cb74 InstanceType=type8 Priority=20
243
INFO[0000] container finished                            ContainerUUID=zzzzz-dz642-000000000000160 State=Complete
244
...
245
INFO[0002] shutdown idle worker                          Age=151.615512ms IdleBehavior=run Instance=stub-providertype8-6ec34c367674cb74 InstanceType=type8 State=idle
246
INFO[0002] instance disappeared in cloud                 Instance=stub-providertype8-6ec34c367674cb74 WorkerState=shutdown
247
</pre>
248 14 Tom Clegg
249 10 Tom Clegg
If the dispatcher starts with a non-empty ARVADOS_DEBUG environment variable, it also prints more detailed logs about other internal state changes, using level=debug.
250
251
h2. Internal details
252
253 38 Tom Clegg
h3. Worker lifecycle
254
255
<pre>
256 41 Tom Clegg
257
  ┌────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
258
  │                                                                                                                                                    │
259
  │                  create() returns ID                                                                                                               │         want=drain
260
  │    ┌───────────────────────────────────────────────────────────────────────────┐                                      ┌────────────────────────────┼─────────────────────────────────────────┐
261
  │    │                                                                           ∨                                      │                            │                                         ∨
262
  │  ┌─────────────┐  appears in cloud list   ┌─────────┐  create() returns ID   ┌─────────┐  boot+run probes succeed   ┌──────┐  container starts   ┌─────────┐  container ends, want=drain   ┌──────────┐  instance disappears from cloud   ┌──────┐
263
  │  │ Nonexistent │ ───────────────────────> │ Unknown │ ─────────────────────> │ Booting │ ─────────────────────────> │      │ ──────────────────> │ Running │ ────────────────────────────> │          │ ────────────────────────────────> │ Gone │
264
  │  └─────────────┘                          └─────────┘                        └─────────┘                            │      │                     └─────────┘                               │          │                                   └──────┘
265
  │                                             │                                                                       │      │                                 idle timeout                  │          │
266
  │                                             │                                                                       │ Idle │ ────────────────────────────────────────────────────────────> │ Shutdown │
267
  │                                             │                                                                       │      │                                                               │          │
268
  │                                             │                                                                       │      │                                 probe timeout                 │          │
269
  │                                             │                                                                       │      │ ────────────────────────────────────────────────────────────> │          │
270
  │                                             │                                                                       └──────┘                                                               └──────────┘
271
  │                                             │                                                                         ∧      boot timeout                                                    ∧
272
  │                                             └─────────────────────────────────────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────┘
273
  │                                                                                                                       │
274
  │   container ends                                                                                                      │
275
  └───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘</pre>
276 38 Tom Clegg
277 10 Tom Clegg
h3. Scheduling policy
278 6 Tom Clegg
279
The container priority field determines the order in which resources are allocated.
280
* If container C1 has priority P1,
281
* ...and C2 has higher priority P2,
282
* ...and there is no pending/booting/idle VM suitable for running C2,
283
* ...then C1 will not be started.
284
285 1 Tom Clegg
However, containers that run on different VM types don't necessarily start in priority order.
286
* If container C1 has priority P1,
287 5 Peter Amstutz
* ...and C2 has higher priority P2,
288 6 Tom Clegg
* ...and there is no idle VM suitable for running C2,
289
* ...and there is a pending/booting VM that will be suitable for running C2 when it comes up,
290 1 Tom Clegg
* ...and there is an idle VM suitable for running C1,
291 6 Tom Clegg
* ...then C1 will start before C2.
292 10 Tom Clegg
293 1 Tom Clegg
h3. Special cases / synchronizing state
294 6 Tom Clegg
295
When first starting up, dispatcher inspects API server’s container queue and the cloud provider’s list of dispatcher-tagged cloud nodes, and restores internal state accordingly.
296 10 Tom Clegg
297 36 Tom Clegg
At startup, some containers might have state=Locked. The dispatcher can't be sure these have no corresponding crunch-run process anywhere until it establishes communication with all running instances. To avoid breaking priority order by guessing wrong, the dispatcher avoids scheduling any new containers until all such "stale-locked" containers are matched up with crunch-run processes on existing VMs (typically preparing a docker image) or all of the existing VMs have been probed successfully (meaning the locked containers aren't running anywhere and need to be rescheduled).
298
299 37 Tom Clegg
At startup, some instances might still be running containers that were started by a prior invocation, even though the (new) boot probe command fails. Such instances are left alive at least until the containers finish. After that, the usual rules apply: if boot probe succeeds before boot timeout, start scheduling containers; otherwise, shut down. This allows the operator to configure a new image along with a new boot probe command that only works on the new image, without disrupting users' work.
300 1 Tom Clegg
301 4 Peter Amstutz
When a user cancels a container request with state=Locked or Running, the container priority changes to 0. On its next poll, the dispatcher notices this and kills any corresponding crunch-run processes (or, if there is no such process, just unlocks the container).
302 6 Tom Clegg
303
When a crunch-run process ends without finalizing its container's state, the dispatcher notices this and sets state to Cancelled.
304 4 Peter Amstutz
305 5 Peter Amstutz
h3. Probes
306
307
Sometimes (on the happy path) the dispatcher knows the state of each worker, whether it's idle, and which container it's running. In general, it's necessary to probe the worker node itself.
308
309
Probe:
310
* Check whether the SSH connection is alive; reopen if needed.
311
* Run the configured "ready?" command (e.g., "grep /encrypted-tmp /etc/mtab"); if this fails, conclude the node is still booting.
312
* Run "crunch-run --list" to get a list of crunch-run supervisors (pid + container UUID)
313 6 Tom Clegg
314 5 Peter Amstutz
h3. Detecting dead/lame nodes
315 10 Tom Clegg
316 28 Tom Clegg
If a node has been up for N seconds without a successful probe, despite at least M attempts, it is shut down, even if it was running a container last time it was contacted successfully.
317
318
h1. Future plans / features
319
320
Per-instance-type VM images: It can be useful to run differently configured/tuned kernels/systems on different instance types, use different ops/monitoring systems on preemptible instances, etc. In addition to a system-wide default, each instance type could optionally specify an image.
321
322 1 Tom Clegg
Selectable VM images: When upgrading a production system, it can be useful to run a few trial containers on a new VM image before making it the default.