On Tue, Feb 18, 2014 at 1:40 PM, Dannon Baker <dannon.baker@gmail.com> wrote:
On Tue, Feb 18, 2014 at 8:30 AM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
This fixed the history_dataset_association.extended_metadata_id error - so is the most likely explanation a failed schema update? Might a stale migration_tmp table have been to blame?
Yes, I've seen this before when I've killed (or otherwise crashed) a migration in process; migrate_tmp doesn't get automatically cleaned up -- and, to allow for recovery, probably shouldn't.
Tricky one - perhaps what would be best is to abort with an explicit error if there is already a (stale) migrate_tmp table present?
Any idea what may have caused it in your case?For a development database I've most commonly just deleted the migrate_tmp table manually and rerun the migration.
I don't know what went wrong here - it could have happened a while back, with the stale migrate_tmp table waiting harmlessly until I ran a schema update today.
It's worth noting that *only* sqlite can have this problem, due to the way migrations work.
I thought that might be the case. Thanks, Peter