Re: ASM setup

On Apr 28, 1:18 pm, macdba321 <macdba...@xxxxxxxxx> wrote:

snip

Thanks for the reply! I actually just got the book you recommended and
since (for once) I have some spare time and some spare hardware, I'm
playing around with it as I read along.

The book is pretty good. Definitely written by the people that
designed the software.


Regardless, since this is all 100% for fun/academic, a quick question:

If you were setting up a server/database for 100% speed (ie, ZERO
requirements for HA and/or disaster recovery .....if I lose a drive,
oh well, I start over, etc.):
(The server has 3 raw disks)
Would you:

A) Hardware-based RAID-0 the 3 disks and let ASM see them afterwards
as 1 "device"
B) Oracle-based RAID-0 the 3 disks
C) Don't use RAID at all
D) Other????

Thanks, and again, this is 100% academic/fun .... this is ZERO
requirement for High Avail!!!!!!!!!!!!!!!!!!!!!!

Vladimir has a good point about the online logs. You can isolate them
on 1 disk ... either an ASM disk or file system based ( if you were
running unix/linux then a direct raw partition is one option ) ... you
can in other words mix and match some parts of the database asm but
others not.

For the other 2 disks why not lie to asm and tell them they are
external redundancy. So create a diskgroup that has just those 2 disk
as external redundancy. Now when you create tablespaces on it asm
will alternate extents between the 2 disks. In other words, you can
use raid 0 to stripe or you can view them as 2 disks and let asm do
that part.

What you really need is about 7 (or more) disks. Use 2 for the online
logs in archive log mode so when you write to one, you then switch
logs, write to the other while the first online log is being archived
to disk 3. Then setup disks 4, 5, 6, 7 as disks in an external
redundancy log group but just allocate part of the disks ( outer
cylinders ) for best IO performance. ( Take your server admin out to
lunch ... by them an itunes gift certificate and bingo ... more
hardware appears eh? ).

.

Relevant Pages

  • Re: ad10: WARNING - READ_DMA UDMA ICRC error (retrying request) LBA=11441599
    ... >>that if I send hardware I want it back when you're done, ... I have two which reliably fail if you put TWO disks on them in a gmirror ... BROKEN by FreeBSD and likely to cause people trouble - including irrevocable ... the machine as the errors cause irrevocable data corruption. ...
    (freebsd-stable)
  • Re: KDE is now broken (Fwd: Heads-up: KDE4 hitting testing tonight (UTC) )
    ... migrate the computer to new hardware. ... Keep it on your disks or put it aside on some removable ... you should probably archive a couple of tape drives along ... Or, at least, an ISA scsi card. ...
    (Debian-User)
  • Re: Choose solaris for ultra10
    ... FreeBSD, and it's really resources hungry, although the hardware are ... I still got stuck to IDE or SCSI disks to boot. ... How it seems to be an general agreement that solaris 10 will run OK, ... Ultra 10 has a notoriously slow IDE ...
    (comp.unix.solaris)
  • Re: Best fs for Oracle RAC
    ... ASM makes migration easier. ... Puget Sound Oracle Users Groupwww.psoug.org ... rebalance to include the EMC disks had completed, ...
    (comp.databases.oracle.server)
  • Re: Emulator or Real Thing ???
    ... Compared to stock hardware, probably not. ... An emulator is only as reliable as the cross-section of the author's coding ... a typical Commodore of any sort, their hard disks will probably die sooner. ...
    (comp.sys.cbm)