PostgreSQL data loads - turn off WAL

Поиск
Список
Период
Сортировка
От hartrc
Тема PostgreSQL data loads - turn off WAL
Дата
Msg-id 1348589351325-5725374.post@n5.nabble.com
обсуждение исходный текст
Ответы Re: PostgreSQL data loads - turn off WAL  (Jayadevan M <jayadevan.maymala@ibsplc.com>)
Re: PostgreSQL data loads - turn off WAL  ("Albe Laurenz" <laurenz.albe@wien.gv.at>)
Re: PostgreSQL data loads - turn off WAL  (Jeff Janes <jeff.janes@gmail.com>)
Список pgsql-general
My version: PostgreSQL v9.1.5
Version string: "PostgreSQL 9.1.5 on x86_64-unknown-linux-gnu, compiled by
gcc (SUSE Linux) 4.3.4 [gcc-4_3-branch revision 152973], 64-bit"

Basically my question is:
Is there currently any way to avoid wal generation during data load for
given tables and then have point in time recovery after that?

Background and blurb
The situation i'm referring to here is for a development environment. I
require point in time recovery because if there is crash etc I don't want to
lose up to a days work for 12 developers. I'm fairly new to PostgreSQL so
please forgive any gaps in my knowledge.

A developer did a data load yesterday of approximately 5GB of data into a
new schema. This generated approximately 7GB of wal.  The situation arises
where if something is incorrect in the data load the data load may need to
be repeated 2 or 3 times (thus generating 20GB +of WAL). For a data load i
don't want wal to be generated. I accept the fact before there was nothing
and from the point of the next pg_basebackup there was everything. It is
from the point i say ok that is everything (the next backup) that i want
point in time recovery to apply to that table.
It is doesn't seem practical, and appears very risky to turn off wal_archive
during the data load.

I'd appreciate your thoughts and suggestions,
Thanks,
Rob





--
View this message in context: http://postgresql.1045698.n5.nabble.com/PostgreSQL-data-loads-turn-off-WAL-tp5725374.html
Sent from the PostgreSQL - general mailing list archive at Nabble.com.


В списке pgsql-general по дате отправления:

Предыдущее
От: Igor Neyman
Дата:
Сообщение: Re: Running CREATE only on certain Postgres versions
Следующее
От: Mike Blackwell
Дата:
Сообщение: Re: unique constraint with significant nulls?