docker service update

Swarm This command works with the Swarm orchestrator.

Description

Updates a service as described by the specified parameters. The parameters are the same as docker service create . Refer to the description there for further information.

Normally, updating a service will only cause the service's tasks to be replaced with new ones if a change to the service requires recreating the tasks for it to take effect. For example, only changing the --update-parallelism setting will not recreate the tasks, because the individual tasks are not affected by this setting. However, the --force flag will cause the tasks to be recreated anyway. This can be used to perform a rolling restart without any changes to the service parameters.

Note

This is a cluster management command, and must be executed on a swarm manager node. To learn about managers and workers, refer to the Swarm mode section in the documentation.

Options

Examples

Update a service

$ docker service update --limit-cpu 2 redis

Perform a rolling restart with no parameter changes

$ docker service update --force --update-parallelism 1 --update-delay 30s redis

In this example, the --force flag causes the service's tasks to be shut down and replaced with new ones even though none of the other parameters would normally cause that to happen. The --update-parallelism 1 setting ensures that only one task is replaced at a time (this is the default behavior). The --update-delay 30s setting introduces a 30 second delay between tasks, so that the rolling restart happens gradually.

Add or remove mounts (--mount-add, --mount-rm)

Use the --mount-add or --mount-rm options add or remove a service's bind mounts or volumes.

The following example creates a service which mounts the test-data volume to /somewhere . The next step updates the service to also mount the other-volume volume to /somewhere-else volume, The last step unmounts the /somewhere mount point, effectively removing the test-data volume. Each command returns the service name.

• The --mount-add flag takes the same parameters as the --mount flag on service create . Refer to the volumes and bind mounts section in the service create reference for details.

• The --mount-rm flag takes the target path of the mount.

$ docker service create \
    --name=myservice \
    --mount type=volume,source=test-data,target=/somewhere \
    nginx:alpine
myservice
$ docker service update \
    --mount-add type=volume,source=other-volume,target=/somewhere-else \
    myservice
myservice
$ docker service update --mount-rm /somewhere myservice
myservice

Add or remove published service ports (--publish-add, --publish-rm)

Use the --publish-add or --publish-rm flags to add or remove a published port for a service. You can use the short or long syntax discussed in the docker service create reference.

The following example adds a published service port to an existing service.

$ docker service update \
  --publish-add published=8080,target=80 \
  myservice

Add or remove network (--network-add, --network-rm)

Use the --network-add or --network-rm flags to add or remove a network for a service. You can use the short or long syntax discussed in the docker service create reference.

The following example adds a new alias name to an existing service already connected to network my-network:

$ docker service update \
  --network-rm my-network \
  --network-add name=my-network,alias=web1 \
  myservice

Roll back to the previous version of a service (--rollback)

Use the --rollback option to roll back to the previous version of the service.

This will revert the service to the configuration that was in place before the most recent docker service update command.

The following example updates the number of replicas for the service from 4 to 5, and then rolls back to the previous configuration.

$ docker service update --replicas=5 web
$ docker service ls
ID            NAME  MODE        REPLICAS  IMAGE
80bvrzp6vxf3  web   replicated  0/5       nginx:alpine

The following example rolls back the web service:

$ docker service update --rollback web
$ docker service ls
ID            NAME  MODE        REPLICAS  IMAGE
80bvrzp6vxf3  web   replicated  0/4       nginx:alpine

Other options can be combined with --rollback as well, for example, --update-delay 0s to execute the rollback without a delay between tasks:

$ docker service update \
  --rollback \
  --update-delay 0s

Services can also be set up to roll back to the previous version automatically when an update fails. To set up a service for automatic rollback, use --update-failure-action=rollback . A rollback will be triggered if the fraction of the tasks which failed to update successfully exceeds the value given with --update-max-failure-ratio .

The rate, parallelism, and other parameters of a rollback operation are determined by the values passed with the following flags:

• --rollback-delay
• --rollback-failure-action
• --rollback-max-failure-ratio
• --rollback-monitor
• --rollback-parallelism

For example, a service set up with --update-parallelism 1 --rollback-parallelism 3 will update one task at a time during a normal update, but during a rollback, 3 tasks at a time will get rolled back. These rollback parameters are respected both during automatic rollbacks and for rollbacks initiated manually using --rollback .

Add or remove secrets (--secret-add, --secret-rm)

Use the --secret-add or --secret-rm options add or remove a service's secrets.

The following example adds a secret named ssh-2 and removes ssh-1 :

$ docker service update \
    --secret-add source=ssh-2,target=ssh-2 \
    --secret-rm ssh-1 \
    myservice

Update services using templates

Some flags of service update support the use of templating. service create for the reference.

Specify isolation mode on Windows (--isolation)

service update supports the same --isolation flag as service create service create for the reference.

Updating Jobs

When a service is created as a job, by setting its mode to replicated-job or to global-job when doing service create , options for updating it are limited.

Updating a Job immediately stops any Tasks that are in progress. The operation creates a new set of Tasks for the job and effectively resets its completion status. If any Tasks were running before the update, they are stopped, and new Tasks are created.

Jobs cannot be rolled out or rolled back. None of the flags for configuring update or rollback settings are valid with job modes.

To run a job again with the same parameters that it was run previously, it can be force updated with the --force flag.