Disaster Recovery Plan

[Xavier Humbert]

Hi,

We plan to proof a DRP for our Rundeck community installation.

* Goal : recover a crashed instance
* Prerequisites : Backup of /etc/rundeck and /var/lib/rundeck/projects.
Dump of the MariaDB database
* Actions : restore data and database, restart rundeckd
* Issues : If Rundeck is in active mode, jobs fire twice, on the
original (still running) instance, and on the DRP instance.
* Question : is the state of Rundeck (active or passive mode) saved in
the database.
* Workaround : with iptables or similar, close outgoing connections on
outgoing port 22

Comments, improvements ?

Thanks, regards,

Xavier

--
Xavier Humbert
CRT Supervision et Exploitation de Niveau 1
Rectorat de Nancy-Metz
03 83 86 27 39

[rac...@rundeck.com]

Hi Xavier,

The state isn’t saved on the database, a good approach is to set rundeck.executionMode=false in the rundeck-config.properties file (as the default state) and then switch to active mode using this API after restoring all instance data.

curl -X POST \
  'http://localhost:4440/api/41/system/executions/enable' \
  --header 'Accept: */*' \
  --header 'X-Rundeck-Auth-Token: WzViUfUnCvXTlWMg1NavSxv3DHHIESfe'

Hope it helps!

​