Re: Disabling WAL for bulk data loads

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



kevinrkeith@xxxxxxxxxxx ("Kevin Keith") writes:
> I am coming from an Oracle background - which in the case of bulk data
> loads there were several options I had where I could disable writing
> to the redo log to speed up the bulk data load (i.e. direct load, set
> the user session in no archive logging, set the affected tables to
> have no logging).
>
> I know the COPY command is one option - however it appears the data
> would need to be in formatted file in order to use it correct? I want
> to avoid writing a new file out for the COPY command and loading that.

The options get a bit better in 8.1, as the COPY command supports some
more format options.

But you'll DEFINITELY get a big boost if you can format things so COPY
can work for you.

> What other options does Postgres 7.4 provide which would allow data
> loads to bypass writing to the WAL? I don't need to have this enabled
> - because in the event of a database crash, I would simply reload the
> data from the source files as recovery.

Arguably you could turn "fsync = off" in postgresql.conf; there is
regrettably the increased risk that a crash would trash the ENTIRE
database, not just the recently-loaded data.

It's somewhat regrettable that you can't turn off logging just for a
single transaction's context; that would probably be the ideal for
this...
-- 
output = ("cbbrowne" "@" "cbbrowne.com")
http://cbbrowne.com/info/spreadsheets.html
"If roach hotels worked on pointy haired people, Microsoft would die."
-- Pete Koren


[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux