You can use the Heirloom Dashboard to roll a new version of your application into production. You might do this because you have made improvements in your application or we have updated the Heirloom platform with new features, applied security fixes, or updated existing systems' runtime environments.
The basic process to create an exact copy of your running production system,
- pick a period of time ("scheduled downtime") when you can make the switch
- take a backup of a running "production" system (with the Instance Portlet)
- restore the backup to a new "staging" instance (with the Backup Calendar Portlet)
- initially the new server and data volume will receive the name "restored-s-xxxx"
- update your application on the "staging" system by deploying a new version of your ElasticCOBOL application onto it or using the file administrator from the home page of your instance (/admin by default)
- carry out testing on the new "staging" system
- transactions against the "staging" will be stored on the new volume
- if customers are still using the "production" system their data may be lost when you switch
- verify that the "staging" system can be terminated & restarted
- at some point in the testing process while the staging system is stopped rename the volume to something that makes sense -- version number, date, etc (Data Volume Portlet)
To place your system into production,
- switch your DNS "alias" record that your customers use to access your application it its new name. If there's a problem,
- switch back
- if the problem was with the platform send us an e-mail to look at during the day
- try again with the same system the next night if there was a problem we fixed or move onto the next system.
- keep the old production system running until you can verify the switch has worked
- terminate your old system
- do not delete the volume as you will lose track of your backups for it
- new backups of the running system will be cataloged under the new volume name
0 Comments