Jon Tirsen
Jon Tirsen
Ah yes I removed the hard coded `us-east-1` and updated the tests. This somewhat breaks backwards compatibility though but probably not in a bad way. Hardcoding `us-east-1` could be disastrous.
I think the best way to implement this is to use FLUSH TABLES WITH READ LOCK on the source mysql database and then stop replication at the correct GTID in...
> @tirsen we think your above proposal is really good! and we only have a few details to append: > > * may need to use `FLUSH TABLES tbl_name [,...
Yeah we only run these diffs against stand by replicas so FTWRL is not a problem. Doesn't chunk checksumming require a special updated_at column maintained by the application? That makes...
Oh I see you just check it a few times until it succeeds? Yeah that might work! That's really nice!
_optimistic data checksum_ sounds like it would certainly work very well for us.
Jib looks pretty awesome. I saw a talk about it on VelocityConf. They could even do debugging and hot code swap inside the docker image. Mind blown.
I started working on extracting the pure JDBC parts from Hibernate to separate it into a `DataSourceModule` (which sets up the Hikari pool and exposes a `DataSource`) and `HibernateModule` (which...
@JGulbronson I'd be happy to review and merge any such PRs though. But start off splitting up the `HibernateModule` since I know almost nothing about SQLDelight except that it's obviously...
Yeah, zipping together two sequences would be very nice. I'd suggest we change this to be a feature request.