Horton

Re: BUG #16604: pg_dump with --jobs breaks SSL connections

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Zsolt Ero <zsolt.ero@gmail.com>
Cc: Magnus Hagander <magnus@hagander.net>, PostgreSQL mailing lists <pgsql-bugs@lists.postgresql.org>, Daniel Gustafsson <daniel@yesql.se>
Date: 2020-09-24T17:42:25Z
Lists: pgsql-bugs 
I wrote:
> Note, however, that AFAICS parallel pg_restore was okay on duplicating
> the original connection parameters otherwise (the prompting issue
> was because it went too far with that...).  So I fail to confirm
> the OP's claim that pg_restore had the same bug as parallel pg_dump.
> If there really is an issue there, we'll need a test case that
> demonstrates it.

Hmm ... so a bit later, I have a test case that may or may not be
what Zsolt saw:

pg_restore -C --dbname="hostaddr=127.0.0.1 sslmode=verify-ca dbname=postgres user=postgres sslcert=client-cert.crt sslkey=client-cert.key" -Fd  dumpdir
pg_restore: error: could not reconnect to database: FATAL:  connection requires a valid client certificate

What is happening here is the exact same issue, but in
ReconnectToServer/_connectDB: those functions suppose that they
can blow away all the original connection parameters and insert
just a database name instead.  I'm inclined to think that
we should get rid of _connectDB, because it seems unnecessarily
duplicative with ConnectDatabase.  But we'll need a bit more
API refactoring than what I had.  New patch coming ...

			regards, tom lane

Commits

  1. Fix connection string handling in psql's \connect command.

  2. Fix connection string handling in src/bin/scripts/ programs.

  3. Fix handling of -d "connection string" in pg_dump/pg_restore.