Releases: vipshop/Saturn
Releases · vipshop/Saturn
Saturn 2.1.5
Saturn 2.1.4
Features
- #191 Provide alarm already read funciton. Once the alarm is marked as already read, then the same alarm will not be raised any more
- #240 Support batch set prefer executor(s), which replace the batch migrate container function.
- #229 Dashboard show overall data.
- #226 Support container group level failover
- #250 Create a page for diff the job configuration data between zk and db.
Improvements
- #248 Make console stateless. Not rely on the session read and write anymore, which make it more container friendly.
- #254 For service downgrade purpose, the console will not refresh the data of zk which not in the same IDC with console.
- #252 As ZK does not support batch read so far, the console will read the configuration data from db instead of zk for improving the speed of page rendering.
- #233 Job list data table support show all data at one page.
Bugfix
- #222 Filter onEnabled and onDisabled events that shoud not be sent to some executors.
- #224 Saturn worker consumes high CPU when Cron describes a past time.
- #255 Register center shouldn't show undefined.
- #253 The perferList information in the ZK is inconsistent with the DB information.
- #220 Rest api, create job, the default value of useDispreferList is not right.
Release Artifacts:
Saturn 2.1.3
Saturn 2.1.2
New Features
- #186 Write job configuration into MYSQL for backup
- #189 Read namespace information from db instead of json file
- #190 Capability to migrate namespace to backup-zookeeper-cluster for disaster recovery
- #194 Let the sharding/dashboard service only involve the zookeeper-cluster that has been configured to avoid cross-IDC calculation
- #185 UpdateCron API
Improvements
- #197 Optimize the Executor deletion process
Bugfix
- #203 RunAtOnce API should check if none of the executors is online, then return error response
- #201 Not check container instance number while it is scaling
- #211 The history job config's table does not correspond to the columns shown
- #204 Shell job return exception during execution
- #207 The job execution page don't show the status
- #200 Fix the "dataTable" column sorting problem
- #195 TimeZone should not be undefined, when console forecast the error cron
- #192 The SignalHandler is not removed, and the loggerContext should not be closed, when embed executor shutdown
- #183 When executor zk reconnected, sharding necessary watcher need be registered, doBusiness need be invoked
Release Artifacts:
Saturn 2.1.0.1
It is a hot fix for 2.1.0
Bugfix
- #182 Saturn Console will force start up the embedded zk during start up
Release Artifacts:
Saturn 2.1.0
New Features
Improvements
- #168 Set timeout for NamespaceShardingService leaderElection
- #169 Add the job number limitation
- #170 Provide command-line options for custom log output paths
- #167 Upgrade jackson version to 2.8.9
Bugfix
- #176 Need retry to connect zk until connected, when zk is lost
- #177 The executor cannot receive the event of RECONNECTED at sometimes
Release Artifacts:
Saturn 2.0.7
New Features
- #113 CreateJob API, GetJobInfo API, RaiseAlarm API
- #144 Provide callback method in AbstractSaturnJavaJob for handling job's onEnabled and onDisabled event
- #149 Provide callback method in AbstractSaturnJavaJob for raising alarm to aslarm endpoint in Console
- #153 Once the SaturnJobReturn contains errorGroup = 551, will raise alarm to alarm enpoint in Console
- #154 Dashboard only be computed in master of console cluster
- #157 Job version support (Java job only)
Improvements
- #142 Add embed demo project
Bugfix
- #121 Saturn plugin ignore current project jar if final name is set in pom.xml
- #147 Once JobSheduler is shutdown, the job business instance should be cleared
- #146 REG_CENTER_JSON_PATH environment prameter should be optional
- #155 Once UncaughtExceptionHandler caught the exception, thread pool should be created a new one
- #158 Executor initialize not correctly when the jobType is 'UNKNOWN-JOB' or in deleteing status.
- #161 js problem when batch remove jobs and jobName contains $ symbol
- #162 NamespaceShardingService doesn't work because of deadlock
Release Artifacts:
Saturn 2.0.6
New Features
- #124 Dashboard support for statistics and alerting.
- #125 Cross time zone job support.
- #117 Rest api support for enable and disable job.
- #134 Batch executor migration support for container based job.
- #111 Job timeout configuraiton support.
Improvements
- #133 reduce the zookeeper watches
Bugfix
- #132 fix the display error of job type when coping jobs
- #136 fix the job didn't run issue when zk-client reconnectd with zk-server between session expire and session lost
- #119 fix the job didn't run issue when deleting the executor node before the sharding result be allocated
- #128 fix the classloader issue in case of starting executor emded
Release Artifacts:
Saturn 2.0.5
New Features
- #40 Integrate with Mesosphere Marathon to support container deployment
Improvements
Bugfix
- #102 fix the thead hang issue when zookeeper connection lost several times in a short period of time
- #107 upgrade fastjson version to avoid security problem
- #109 fix failover issue when disabling the job
- #106 fix concurrency issue of sharding content
- #103 fix event lost issue during gray upgrade process
Release Artifacts:
Saturn 2.0.4
New Features
- #91 add group attribute for jobs which can be used for group filter
- #90 add job dependency attribute for jobs and do dependency check when starting/stopping jobs
Improvements
- #92 add Executor version information in the registry page
Bugfix
- #96 fix ConcurrentModificationException when stopping Executor
- #83 fix IllegalMonitorExcetion when handling job timeout strategy
- #94 fix concurrency issue of sharding content when enabling hundreds of jobs