Bug #51652 ndb_restore needs a --force option
Submitted: 2 Mar 2010 18:10
Reporter: Andrew Hutchings Email Updates:
Status: Verified Impact on me:
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S4 (Feature request)
Version:mysql-5.1-telco-6.2 OS:Any
Assigned to: Assigned Account CPU Architecture:Any

[2 Mar 2010 18:10] Andrew Hutchings
A --force option should be added to ndb_restore which will ignore errors when restoring schemas and data (but still report them).

How to repeat:
[8 Mar 2010 14:20] Jørgen Austvik
Is force here "ignore all errors", or "overwrite". Are there any errors or conditions it should stop on even with --force? - we might need more requirements.
[8 Mar 2010 14:26] Andrew Hutchings
When I started looking through this properly over the weekend I came up with similar questions.

My idea is for it to be a little like doing "mysql --force < dump.sql".  In that errors are ignored (where possible).

In addition to this it should remove the table from the list in memory so it is not used as part of the next step of the restoration.

Finally as a bonus I was considering a table already exists error on a metadata restore to compare with the table data in the .ctl to see if the schema matches (not looked into how easy this bit is to do yet).

Having said all this the default action for a data restore is to silently overwrite, so maybe that should be the behaviour?