Mladen Gogala wrote:
> On Fri, 24 Jun 2005 13:09:30 -0700, DA Morgan wrote:
>
>
>>This is more a matter of where the space is, how you plan to back it
>>up, and the rest of your storage strategy and hardware.
>
>
> Nope, not true. It depends on the volume of data. Slow archiving can
> cause "checkpoint not complete" errors and can, in extreme cases, hang
> your system. Also, both nodes need to see all archives, but they don't
> need to lock them. That allows NFS. If you have 1GB redo logs and are
> going through that once a minute because of,say, a huge data load which
> for some reason cannot be done in nologging mode, you better make sure
> that your OS is capable of writing more then a gig per minute. Having
> archives on NFS drive is an exceptionally bad idea if the server goes
> offline. Then your database will be hung by the neck until the server is
> back up. May God have mercy on your database.
I agree completely but it would be the rare local drive that was faster
than the cache in a disk array.
--
Daniel A. Morgan
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)
Received on Sat Jun 25 2005 - 10:19:59 CDT
