Thank you for responding. I did see that note and should have included that as 
part of my question.

Since pg_upgrade is in control of how it is calling pg_dump, is there a reason 
pg_upgrade cannot use the directory output format when calling pg_dump? Is the 
schema-only operation incompatible?

________________________________________
From: Adrian Klaver <adrian.kla...@aklaver.com>
Sent: Saturday, April 6, 2019 1:52 PM
To: senor; pgsql-general@lists.postgresql.org
Subject: Re: pg_upgrade --jobs

On 4/6/19 11:44 AM, senor wrote:
> The pg_upgrade --jobs option is not passed as an argument when it calls 
> pg_dump. I haven't found anything in docs or forums mentioning a reason for 
> not supporting under certain circumstances other than possibly for pre-9.2. 
> The pg_upgrade docs page states that it allows multiple CPUs to be used for 
> dump and reload of schemas. Some databases I'm upgrading have 500,000+ tables 
> and running with a single process is greatly increasing the upgrade time.
>
> I am also using the --link option.
> I have tried "--jobs 20", "--jobs=20", placing this option first and last and 
> many other variations.
> I am upgrading 9.2.4 to 9.6.12 on CentOS 6.
> Varying hardware but all with 32+ CPU cores.
>
> su - postgres -c "/usr/pgsql-9.6/bin/pg_upgrade --jobs=20 --link \
>         --old-bindir=/usr/pgsql-9.2/bin/ \
>         --new-bindir=/usr/pgsql-9.6/bin/ \
>         --old-datadir=/var/lib/pgsql/9.2/data/ \
>         --new-datadir=/var/lib/pgsql/9.6/data/"
>
> I feel like there's a simple reason I've missed but this seems pretty 
> straight forward.

https://www.postgresql.org/docs/9.6/app-pgdump.html

"--jobs=njobs

     Run the dump in parallel by dumping njobs tables simultaneously.
This option reduces the time of the dump but it also increases the load
on the database server. You can only use this option with the directory
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
output format because this is the only output format where multiple
^^^^^^^^^^^^^
processes can write their data at the same time."


> A secondary plan would be to find instructions for doing the same as 
> "pg_upgrade --link" manually so I can run "pg_dump --jobs 20".
> Any assist is appreciated.
> Thanks,
> S. Cervesa
>
>


--
Adrian Klaver
adrian.kla...@aklaver.com


Reply via email to