pg_dump.exe compared to MS SQL backup

Question

One of my clients recently migrated a CRM product from MS SQL to PostgreSQL. I have a question about how data backup works using pg_dump.exe

As so far in MS SQL, when the backup started, the database was given a "time stamp" and the database copy was performed at a given moment, regardless of whether new data was added to the database during, say, an hour's data backup. Just a copy and so it did for the selected initial moment of time, and the copy was coherent for that moment.

At least that's how I understand the process and how I can describe it. You can verify my knowledge if I'm wrong. If I am correct, please let me know what this "ability" is called.

Now let's move on to PostgreSQL. Is the data copy done the same?

I mean, if the data backup process starts at midnight and takes about a hour, is it possible to start working with CRM, for example, a minute after starting the backup, or do you have to hold off on making changes until the backup is completed?

Answer 1

The data produced by pg_dump corresponds to the state of the database at the start of the backup, and is not affected by changes occuring during the backup. From pg_dump's documentation:

It makes consistent backups even if the database is being used concurrently. pg_dump does not block other users accessing the database (readers or writers).

The reason is that the complete dump is performed in a single REPEATABLE READ transaction.