From: | Stephen Frost <sfrost(at)snowman(dot)net> |
---|---|
To: | Thorsten Schöning <tschoening(at)am-soft(dot)de> |
Cc: | pgsql-admin(at)lists(dot)postgresql(dot)org |
Subject: | Re: What's the dfifference between pg_start_backup+copy+pg_stop_backup+WAL vs. pg_start_backup+pg_stop_backup+copy+WAL? |
Date: | 2021-07-23 19:22:12 |
Message-ID: | 20210723192212.GO20766@tamriel.snowman.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
Greetings,
* Thorsten Schöning (tschoening(at)am-soft(dot)de) wrote:
> Guten Tag Scott Ribe,
> am Freitag, 23. Juli 2021 um 20:07 schrieben Sie:
>
> > Because it is in the file. WAL can re-do changes, it cannot undo them.
>
> But isn't visibility of rows associated to transaction numbers as
> well? I would have expected the transaction number for the row in B
> simply not being available, because it was never part of the WALs up
> to and including pg_stop_backup when not following the official
> ordering of commands. So I expected that row to simply not be visible
> at all and vacuumed at some point or whatever.
The information about what transactions have committed and what haven't
isn't just in the WAL- it's in the tuple header for frozen tuples, or
all visible tuples, and it's in the transaction log files otherwise, so
it would depend on when those are copied.
Thanks,
Stephen
From | Date | Subject | |
---|---|---|---|
Next Message | Stephen Frost | 2021-07-23 19:23:12 | Re: What's the dfifference between pg_start_backup+copy+pg_stop_backup+WAL vs. pg_start_backup+pg_stop_backup+copy+WAL? |
Previous Message | Thorsten Schöning | 2021-07-23 19:15:17 | Re: What's the dfifference between pg_start_backup+copy+pg_stop_backup+WAL vs. pg_start_backup+pg_stop_backup+copy+WAL? |