The evidence is that he didn't read the postgres manual section on log-based replication[1] which would have told him how to configure a postgres master server so that it doesn't delete logs until all consumers have processed them.
It's not a five minute setup, but Dan doesn't write that the setup takes longer than five minutes - he writes that the design is fundamentally broken. Which it isn't, if you read the postgres manual. We're not even talking about the manual of the product he tried out for five minutes - we're talking about the manual of the database he's responsible for administering!
The overall point of the article is fine though. Original Commenter was nitpicking.
> how to configure a postgres master server so that it doesn't delete logs until all consumers have processed them.
Do you think it'd be reasonable for FiveTran to include this little tidbit in their setup documentation? I'm not talking about repeating the Postgres docs, but just a blurb about the need to do this kind of Postgres config?
That's an example of what I mean when I'm calling out georgewfraser to be humble and use Dan's feedback improve his product (in this case by improving the docs) instead of name calling his customers.
Ok, so Dan came to the wrong conclusion and was wrong to say the product was broken, but he had the professional courtesy to not name the company/product. George just attacks his character. Like another commenter mentioned, we don't even know for sure it was FiveTran. Yet, George just jumps in head first with guns blazing.
It's not a five minute setup, but Dan doesn't write that the setup takes longer than five minutes - he writes that the design is fundamentally broken. Which it isn't, if you read the postgres manual. We're not even talking about the manual of the product he tried out for five minutes - we're talking about the manual of the database he's responsible for administering!
The overall point of the article is fine though. Original Commenter was nitpicking.
[1] https://www.postgresql.org/docs/current/warm-standby.html