项目作者: xebialabs-community

项目描述 :
The Lock plugin is a XL Deploy plugin that adds capabilities for preventing simultaneous deployments.
高级语言: Java
项目地址: git://github.com/xebialabs-community/xld-lock-plugin.git
创建时间: 2014-09-01T15:42:00Z
项目社区:https://github.com/xebialabs-community/xld-lock-plugin

开源协议:MIT License

下载


XLD Lock plugin

Build Status
License: MIT
Github All Releases

Preface

This document describes the functionality provided by the XLD Lock plugin.

See the XL Deploy reference manual for background information on XL Deploy and deployment concepts.

Overview

The XLD Lock plugin is a XL Deploy plugin that adds capabilities for preventing simultaneous deployments.

Features

  • Lock a specific environment / application combination for exclusive use by one deployment
  • Lock a complete environment for exclusive use by one deployment
  • Lock specific containers for exclusive use by one deployment
  • List and clear locks using a lock manager CI
  • Wait for lock

Requirements

  • XL Deploy requirements
    • XL Deploy: version 4.0+
    • Other XL Deploy Plugins: None

Installation

  • Place the plugin JAR file into your SERVER_HOME/plugins directory.
  • Restart the XL Deploy server.

Build it

Following options are available:

  • gradle clean assemble: Will generate a jar that can be installed.
  • gradle clean test: Will execute the unit tests (if any)
  • gradle clean itest: Will execute the integration tests (XLDEPLOY_HOME environment variable to be set). For example: XLDEPLOY_HOME=/opt/xldeploy/xl-deploy-4.0.1/xl-deploy-4.0.1-server

Usage

Locking deployments

When a deployment is configured, the Lock plugin examines the CIs involved in the deployment to determine whether any of them must be locked for exclusive use. If so,
it contributes a step to the beginning of the deployment plan to acquire the required locks. If the necessary locks can’t be obtained, the deployment will enter a PAUSE
state and can be continued at a later time. If the enviroment to which the deployment is taking place has its enableLockRetry property set, then the step will wait for a period of time before retrying to acquire the lock.

If lock acquisition is successful, the deployment will continue to execute. During a deployment, the locks are retained, even if the deployment fails and requires
manual intervention. When the deployment finishes (either successfully or is aborted), the locks will be released.

Configuration

The locks plugin adds synthetic properties to specific CIs in XL Deploy that are used to control locking behavior. The following CIs can be locked:

  • udm.DeployedApplication: this ensures that only one depoyment of a particular application to an environment can be in progress at once
  • udm.Environment: this ensures that only one depoyment to a particular environment can be in progress at once
  • udm.Container: this ensures that only one depoyment can use the specific container at once

Each of the above CIs has the following synthetic property added:

  • allowConcurrentDeployments (default: true): indicates whether concurrent deployments are allowed. If false, the Lock plugin will lock the CI prior to a deployment.

The udm.Environment has the following additional synthetic properties :

  • lockAllContainersInEnvironment (default: false) If set, will lock all containers in environment instead of only the enviroment
  • enableLockRetry (default: false): If set, will not PAUSE the deployment on failure to acquire locks. Instead continually tries to obtain the lock after a period of time.
  • lockRetryInterval (default: 30): Seconds to wait before retrying to obtain lock.
  • lockRetryAttempts (default: 60): Number of retry attempts. On failure to obtain locks after the designated attempts, the deployment will be PAUSED.

Implementation

Each lock is stored as a file in a directory under the XL Deploy installation directory. The lock.Manager CI can be created in the Infrastructure section of XL Deploy to list and clear all of the current locks.