We had a bug in our data copying script, which didn't reset the ID sequences properly. This has been fixed in 1.1.30. To deploy, we should: * drop existing RDS campaign instance * drop the alembic_version table in Redshift * run `SPLICE_IGNORE_REDSHIFT=true manage.py db upgrade` * run the rs_to_pg.py script * test deploying a past distribution
This works, but: Also creates the alembic_version table in redshift. Probably doesn't matter, but worth mentioning: throughout the code `if 'SPLICE_IGNORE_REDSHIFT' in os.environ` is used, which means that SPLICE_IGNORE_REDSHIFT=false would do the wrong thing.
Hey Daniel, Thanks for the updates. Yes, alembic always creates "alembic_version" for version control. Can't ignore this one. The "SPLICE_IGNORE_REDSHIFT" was initially used as a flag, doesn't matter what value it is. Will fix it, make it explicit. Can I use staging Splice to do some tests now? Wonder how much faster the traffic is now. Thanks,
yep, go ahead and test away