HA Account Promotion

Contents

HA Account Promotion#

The Active Replica mechanism underlying Carbonio HA is described in Section Carbonio HA. In particular, Replica Promotion shows how to manually activate a Replica. To automatise this process, the habeat Python tool has been developed to ensure seamless account promotion with high availability.

The habeat tool can be downloaded from the zextras/sps-habeat repository. You will need also a number of other files from that repository, so you might want to clone it.

All the commands in this section must be executed as the root user for every HA Node, i.e. for every Node listed in the column HA Nodes in Table 7. Taking into account our inventory file, this means you must install and configure the utility on these Nodes:

Table 8 Nodes and FQDN#

HA Node

FQDN

MTA

mta2.example.com

Proxy

proxy2.example.com

Mailstore & Provisioning

mbox2.example.com

Collaboration

filesdocs2.example.com

Video Server

video2.example.com

Deploy#

To copy habeat you can use preferable utility, for example scp. Remember to replace node with the actual Node FQDN as shown in Table 8 or with equivalent FQDN according to your infrastructure.

First, copy the script and its configuration file.

Hint

Before copying the configuration file, you might want to edit it to adapt it to your infrastructure. Please refer to Section Configure below.

# scp habeat.py root@node:/usr/local/sbin/habeat.py
# ssh root@node "mkdir -p /etc/hamon"
# scp config/habeat.yml root@node:/etc/hamon

To configure habeat you need to add 2 units to systemd service:

  • habeat.service – the service that should perform call of script

  • habeat.timer – the timer for define how often to run the script

Copy them to each Node.

# scp config/habeat.service root@node:/etc/systemd/system/
# scp config/habeat.timer root@node:/etc/systemd/system/
# ssh root@node "systemctl daemon-reload"

In the habeat.timer unit we can define the condition of execution, i.e., the interval between each script execution.

OnCalendar=\*:0/5 # Run script every 5 minutes

In the habeat.service unit we define the location of the script, log file, and configuration file.

ExecStart=/usr/local/sbin/habeat --config /etc/hamon/habeat.yml --log /var/log/habeat.log

Configure#

The configuration file, which you can find also in the repository, is similar to the following: as usual, remember to fill the options with values suitable to your infrastructure.

Habeat configuration file
local:
  whoami: "secondary" # marker on which dc script is running
  dc_check: "primary" # marker which dc script should check
  role: "appserver" # supported roles: appserver or consulserver or proxyserver
  checkDownFile: "/var/tmp/appserverdown-habeat" # template for lock file related with down state
  checkPromotionFile: "/var/tmp/appserverprom-habeat" # template for lock file related with promotion state
  checkRestartReplicaFile: "/var/tmp/appserverrstopha" #only app server lock file for restart replica lock file
  provider: consul # provider for external witness: hetrix or vcenter
  threads: 5 # carbonio ha promotion number of threads
  proxy_switch: "bgp" # proxy switch tool: bgp
  proxy_enable_activate: false # enable run proxy switch
  proxy_enable_deactivate: false # enable run proxy reverse switch
  disable_ha_module: false # define if another app server down stop ha module of Carbonio
  flush_cache: false # define if we need flush cache for accounts after promotion
  flush_arguments_a: false # define if we need run flush cache for all application services. Used with flush_cache: true
  restart_replica: false # run ha restartReplicas accounts

primary:
  proxy_ip: # this value used for check availability of proxy in primary dc
  appserver_ip: delete # this value used for check availability of application in primary dc
  directorysrv_ip:
  consul_ips: #this values used for check availability of consul servers in primary dc
  -
  -
  consul_vmnames: # this values used for check consul servers in monitoring server in primary dc
  - svc1.example.com
  - svc3.example.com
  appserver_vmname: mbox1.example.com # this value used for check application server in monitoring server in primary dc
  proxyserver_vmname: proxy1.example.com # this value used for check proxy server in monitoring server in primary dc

secondary:
  proxy_ip: # this value used for check availability of proxy in secondary dc
  appserver_ip: # this value used for check availability of application in secondary dc
  directorysrv_ip:
  consul_ips: # this values used for check availability of consul servers in secondary dc
  -
  consul_vmnames: # this values used for check consul servers in monitoring server in secondary dc
  - svc2.example.com
  appserver_vmname: mbox2.example.com # this value used for check application server in monitoring server in secondary dc
  proxyserver_vmname: proxy2.example.com # this value used for check proxy server in monitoring server in secondary dc
  proxy_switch:
    bgp: # this value used for choose proxy switch
      activate: # list of command to activate proxy switch
      deativate: # list of command to deactivate proxy switch

consul: # consul provider config
  hostname: 127.0.0.1
  port: 8500
  token:
Contents