Skip to content

pg_backup_start #5

@nmoreaud

Description

@nmoreaud

Hello

I am testing the concept of this extension, manually.
On my local database, if I modify a table and then make a copy of pgdata/base/XXX, the copy doesn't contain the modified data.
I need to call pg_backup_start('test', True) and keep the session open during the copy to ensure the data is flushed into pgdata/base/XXX directory.

It also seems that using subvolumes is not required.
This command works well:
cp -ra --reflink=always pgdata/base/159591 pgdata/base/111111

I'm not sure if this technique can be used for production software.
PostgreSQL documentation contains a lot of warnings related to filesystem copies: https://www.postgresql.org/docs/current/backup-file.html

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions