View Source

When trying to restore a MySQL DB with InnoDB tables in CDP, you receive the following error:
[2010-Dec-28 22:06:35],[ WARN],Could not stop the mysqld instance
[2010-Dec-28 22:06:35],[ERROR],An exception occurred during the request. Could not ping the secondary mysql instance (after starting the instance)
[2010-Dec-28 22:06:35],[DEBUG],failed to prepare mysql instance
When initiating a DB restore, the CDP Server will transfer MySQL data (InnoDB tablespace) to the Agent, that will write the data into the cdp-agent's tmp directory ({color:blue}/var/lib/r1soft/tmp{color}). The CDP Agent will also create a temporary instance of MySQL to conduct the restore to. This is where the problem is occurring. The agent is unable to start the temporary instance. Common causes for this failure are security mechanisms such as SELinux and Apparmor.
If caused by SELinux, put SELinux in permissive mode while doing the restore. If running Apparmor, run "{color:#0000ff}service apparmor teardown{color}", simply running '{color:blue}aa-complain{color}' for mysqld and "{color:blue}service apparmor stop{color}" will not resolve the issue. Once a tmp MySQL instance is created, the Connector/J will connect to both instances and migrate the InnoDB database from the {color:blue}tmp{color} instance to the primary instance.