v1.2.4-rc1

1.2.4-rc1 (2016-06-14)

API

• Code refactoring

Test

• Fix leader election tests

v1.2.3

1.2.3 (2016-05-27)

API

• Update engine-api vendoring (supports new functions and signatures)
• Fix registry auth bug for image pulls

v1.2.3-rc1

1.2.3-rc1 (2016-05-25)

API

• Update engine-api vendoring (supports new functions and signatures)
• Fix registry auth bug for image pulls

v1.2.2

1.2.2 (2016-05-09)

Cluster management

• Fix deadlock that causes Swarm to hang

v1.2.2-rc1

1.2.2-rc1 (2016-05-05)

Cluster management

• Fix deadlock that causes Swarm to hang

v1.2.1

1.2.1 (2016-05-03)

Scheduler

• Add containerslots filter to allow user to limit container number on a node

API

• Fix and document some api response status codes
• Add NodeID in docker info
• Support docker ps --filter by volume

Build

• Revert to Go 1.5.4
• Convert docs Dockerfiles to use docs/base:oss

Test

• Fix api/ps tests
• Update api/stats test to prevent timeout on master branch
• Use --cpu-shares instead of -c in integration test

Misc

• Close http response body to avoid potential memory leak
• Switch context.TODO() to context.Background() to enable context setting

v1.2.1-rc1

1.2.1-rc1 (2016-04-22)

API

Use engine-api to handle large number of API calls
Update ContainerConfig to embed HostConfig and NetworkingConfig
stop/restart/kill an non-existent container should returns 500 rather than 404
Return an error when assertion fails in hijack
Return error when Image Pull fails
Fix image pull bug (wait until download finishes)

Build

Switch to go 1.6.2
Move dependencies to vendor/
Updating Image Pull to use docker/distribution package

Misc

Documentation clean up

v1.2.0

1.2.0 (2016-04-13)

Scheduler

• Move rescheduling out of experimental
• Differentiate constraint errors from affinity errors
• Printing unsatisfiable constraints for container scheduling failure
• Enable rescheduling on master manager to prevent replica managers from rescheduling containers
• Output error when starting a rescheduled container fails, and when removing container fails at node recovery
• Validate cluster swarm.overcommit option

API

• Introduce engine-api client to Swarm
• Implement 'info' and 'version' with engine-api
• Use apiClient for some volume, network, image operations
• Print engine version in Info
• Fix swarm api response status code
• Support ps node filter
• Fix HostConfig for /start endpoint
• Print container 'created' state at ps
• Update dockerclient to get labels on volumes, networks, images
• Support private images, labels and other new flags in docker build
• Select apiClient version according to node docker version

Node management

• Prevent crash on channel double close
• Manager retries EventMonitoring on failure.
• Docker engine updates hostname/domainname
• Force inspect for containers in Restarting state.
• Increase max thread count to 50k to accommodate large cluster or heavy workload
• Force to validate min and max refresh interval to be positive
• Skip unstable tests from Docker bug 14203
• Fix race condition between node removal from discovery and scheduler
• Fix data race with node failureCount
• Display warning message if an engine has labels with "node=xxx"

Discovery

• Remove parameter which is not used in createDiscovery
• Fix Consul leader election failure on multi-server

Mesos integration

• Support rescind offer in swarm
• Update mesos tests

Misc

• Update golang version to 1.5.4
• Skip redundant endpoints in "network inspect"
• Validate duration flags:--delay, --timeout, --replication-ttl
• Fix image matching via id
• Make port 0 invalid as listening port
• Improve volume inspect test
• Add read lock for eventsHandler when only it is necessary

v1.2.0-rc3

1.2.0-rc3 (2016-04-08)

Scheduler

• Report error when removing container fails at node recovery
• Validate cluster swarm.overcommit option

API

• Select apiClient version according to node docker version

Node management

• Fix race condition between node removal from discovery and scheduler
• Fix data race with node failureCount
• Display warning message if an engine has labels with "node=xxx"

Discovery

• Fix Consul leader election failure on multi-server

v1.2.0-rc2

1.2.0-rc2 (2016-04-01)

Scheduler

• Enable rescheduling on master manager to prevent replica managers from rescheduling containers
• Output error when starting a rescheduled container fails

API

• Use apiClient for some volume, network, image operations
• Print engine version in Info
• Print container 'created' state at ps
• Update dockerclient to get labels on volumes, networks, images
• Support private images, labels and other new flags in docker build

Misc

• Make port 0 invalid as listening port
• Improve volume inspect test
• Skip unstable tests from Docker bug 14203
• Add read lock for eventsHandler when only it is necessary