| From: | Greg Sabino Mullane <htamfids(at)gmail(dot)com> |
|---|---|
| To: | KK CHN <kkchn(dot)in(at)gmail(dot)com> |
| Cc: | pgsql-general(at)lists(dot)postgresql(dot)org |
| Subject: | Re: Pgbackrest info output interpretation |
| Date: | 2025-11-14 13:48:57 |
| Message-ID: | CAKAnmmLUE=BmgBbkqFp9nuQd95PxHAxx=R3Q_WH2YLT_vYg4YA@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
The "database size" is going to be your data directory, MINUS your pg_wal
directory, MINUS unlogged/temp relations, and MINUS things that are not
needed for recovery (i.e. all the small pg_ directories such as
pg_subtrans). The "database backup size" is how much of that 3GB is part of
*this* backup - for a full backup, the number will be the same, for diff or
incr, it will be a lot less. The "repo backup size" on the last line is the
compressed size of the previous "database backup size"
How to correlate the pgbackrest ( info) database size, database backup size
> and repo1 backup size and actual du output on DB cluster.
>
You cannot, really, without deep knowledge of things like which files on
disk map back to unlogged tables. However, du on the datadir minus all pg_
dirs should get you in the ballpark. As a quick example:
cd $DATADIR
du --summarize * | awk '!/pg_/{x=x+$1}END{print x}' | numfmt --to=iec
Cheers,
Greg
--
Crunchy Data - https://www.crunchydata.com
Enterprise Postgres Software Products & Tech Support
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Marcel Menzel | 2025-11-14 18:16:16 | pg_upgrade reflink support on OpenZFS |
| Previous Message | Ron Johnson | 2025-11-14 13:42:46 | Re: Pgbackrest info output interpretation |