Re: ASM LUN sizes and number of disks
- From: "Greg Rahn" <greg@xxxxxxxxxxxxxxxxxx>
- To: finn.oracledba@xxxxxxxxx
- Date: Fri, 7 Nov 2008 20:36:19 -0800
On Fri, Nov 7, 2008 at 6:55 PM, Finn Jorgensen <finn.oracledba@xxxxxxxxx> wrote:
> Because of the way ASM distributes data evenly across all disks in a
> diskgroup, adding 200GB disks to an existing diskgroup comprising of 50GB
> disks means you will never be able to use more than 50GB of those 200GB
> disks.
I agree with this...
> You will have to add those disks to a separate diskgroup and then start
> moving data over, which means downtime.
but could you not add(2x200)/rebalance, drop(6x50GB)/rebalance to get
all the data onto the 2x200GB?
this would alleviate downtime but...
I would never put my whole database only on 2 spindles anyway. If I
recall correctly, if there is not room to mirror ASM extents from the
failed drive, the diskgroup will dismount to protect from catastrophic
failure. This means in a two disk ASM group, losing one disk will
result in a down database. There there is also a special case with 3
disks, as each is in their own failgroup and I believe the usable
space (for files) for 3 disks is the same as 2 disks (or something
similar). So basically the minimum recommended disks is four. It
comes back to disk space is cheap...
--
Regards,
Greg Rahn
http://structureddata.org
--
//www.freelists.org/webpage/oracle-l
Other related posts:
