| From: | Timothy Garnett <tgarnett(at)panjiva(dot)com> |
|---|---|
| To: | sthomas(at)optionshouse(dot)com |
| Cc: | Joachim Wieland <joe(at)mcknight(dot)de>, Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc>, Andrew Dunstan <andrew(at)dunslane(dot)net>, Dimitri Fontaine <dimitri(at)2ndquadrant(dot)fr>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Allowing parallel pg_restore from pipe |
| Date: | 2013-05-16 17:16:17 |
| Message-ID: | CAPcyiQ0GZMgFJvbXG=DCVktj=MgcdDNSFoLMxNuRyYOa6S6e1g@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> If you need something like this short term, we actually found a way to do it
> ourselves for a migration we performed back in October. The secret is xargs
> with the -P option:
>
> xargs -I{} -P 8 -a table-list.txt \
> bash -c "pg_dump -Fc -t {} my_db | pg_restore -h remote -d my_db"
>
> Fill table-list.txt with as many, or as few tables as you want. The above
> example would give you 8 parallel threads. Well equipped systems may be able
> to increase this.
>
> Admittedly it's a gross hack, but it works. :)
I think you'd have to be real careful around foreign key constraints
for that to work.
Tim
| From | Date | Subject | |
|---|---|---|---|
| Next Message | David E. Wheeler | 2013-05-16 17:18:19 | Re: Patch proposal: query result history in psql |
| Previous Message | Tom Lane | 2013-05-16 17:15:42 | Re: Heap truncation without AccessExclusiveLock (9.4) |