Alexander Kuzmenkov

The entire application might be a little too much for us to go through. To better investigate your problem, let me describe how I understand the issue. You create an...

``` ereport(ERROR, (errcode_for_file_access(), errmsg("could not close pipe to external command: %m"))); ``` Postgres wants to keep its mysteries and doesn't tell us which program it had problems with... I think...

I'd say that a program error like this is definitely a bug and probably high priority. Good that we have a workaround, bad that we don't know how to reproduce...

The usual mystery disappearance of the `disttable_repl`... I added statement logging to investigate it, does the server log show anything?

With LATERAL join on timestamp inequality, I guess the right side is going to do index lookups, but it's still O(left_rows * log(right_rows)). A merge-join-like algorithm will be O(left_rows +...

The different grammar is probably required because the semantics is different from the normal join on inequality condition -- we only have to return the closest righthand row that matches....

Wrote a design memo for internal use: https://docs.google.com/document/d/1YEX038V-gq-iLfM-KpLV0zpNn3yW88I9qPVa0_bs4aQ/edit

The discussion upstream: https://www.postgresql.org/message-id/flat/CALzhyqwuVz0FJZ-oCYQ9d%2ByrPrbF5a9HDyAjxuSUdgq8n7nshQ%40mail.gmail.com

We could also make `elog` abort in debug. I did a similar thing for some other DBMS and it was very efficient in enforcing separation between "internal program logic errors"...

I think this is kinda bug-ish, ideally we shouldn't require the user to work around our suboptimal scheduling. I imagine the slowdown is due to IO concurrency, so we could...