Rolling back from a failed version upgrade

[shaerul]

I’ve searched extensively online for proper documentation on rolling back management servers and hosts after a failed upgrade but couldn’t find anything suitable. Can anyone share a reliable, proven, step-by-step process? Thanks.

[pcfriek1987]

I don't think it's that easy with database schema updates etc, you might want to explain in more detail what went wrong.

[weizhouapache]

If you have a db dump (by mysqldump -R) of cloud and cloud_usage databases, you just need to

• stop cloudstack-management
• restore mysql databases from backup
• restore /etc/cloudstack/management/ (sometimes it is needed, for example 4.19->4.20 upgrade which include log4j upgrade)
• downgrade cloudstack-management , cloudstack-common and cloudstack-usage package
• restart cloudstack-management

I did it many times without issue.