Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Oracle Disk Configuration question.

Re: Oracle Disk Configuration question.

From: Paul Drake <drak0nian_at_yahoo.com>
Date: 1 Jun 2004 22:33:33 -0700
Message-ID: <1ac7c7b3.0406012133.8ea057b@posting.google.com>


bc9am <bc9am_at_hotmail.com> wrote in message news:<ph7vc.248$Xh3.65_at_newsfe1-gui.server.ntli.net>...
> Hi,
>
> I'm trying to understand what the best disk configuration is for an
> Oracle server I am trying to build. Currently the server is spec'ed to
> have 4 disks.

There you have it. 4 disks it is.

> In my mind that gives me two choices. Either two sets of
> RAID 1, or a big RAID 5 array.

or you ask for more disks (and a sufficient number of controllers/channels).
RAID 10 should also be a consideration.

> It's a high availability server so I
> can't afford downtime due to a dead disk.

good, that sounds like a design spec, a much better one than "the server shall have 4 disks". If its truly high availability, you will need other measures in place besides having a fault tolerant storage configuration, such as hot spares, a disaster recovery server, standby database, RAC config. As others here will likely mention, a well tested and documented set of procedures for install, backup/restore/recover is an excellent (and lowest cost) high availability "solution", second only to a well trained and competent staff. Get the simple stuff down first.

Objective: fault tolerant storage configuration. A single disk failure should not cause loss of data. System Failure and/or loss of data due to loss of a RAID controller, mainboard, power supply, ethernet adapter, memory module will be tolerated.

this is going to be attached to a UPS, and you'll have a shutdown script that will be triggered by the UPS, right? you'll also have a battery backup on the RAID controller, right? you're much more likely to lose power than to lose a hard drive. Also - if the air conditioning in the server room is lost, an alarm will page staff? If you lose air conditioning, you'll likely lose all of your hard drives, if the server does not shut itself down.

> After a bit of research I have understood that you are supposed to put
> all your core database (table?) files (and any random access stuff [inc
> OS files?]) on one disk, and sequential stuff on their own disks (e.g.
> logs).

not a bad rule of thumb, but there are many such rules or heuristics. The concepts guide tends to discuss these pretty well. Typically, one strives to match file block sizes and read/write sizes to filesystem, operating system and RAID volume sizes (or multiples thereof).
How did you choose the database block size?

> I obviously don't have enough disks to put all sequential files
> on their own disks. So, with this in mind I guess logic suggests putting
> them all on one disk.

how about we define the terms:
 "disk" = physical hard drive
 "RAID volume" = striped and or mirrored set of physical disks, need not entirely consume those physical hard drives.  "partition" = contiguous section of storage, either on a single "disk" or on a "RAID volume". need not necessarily support a filesystem.
 "filesystem" = formatted partition.

and go from there.

> However this seems a bit silly as then its just
> random access (?) - so I might as well have a RAID 5 array for the whole
> lot?

So far you have not mentioned storage capacity requirements. Why not just use a single RAID 10 volume if the RAID controller supports it?
(and create partitions, which you can format with filesystems of the filesystem type, block size and size of your chosing, which will be mapped to mount points such as /, /usr, /home, /opt, <swap> ...).

> Obviously with this particular configuration cost is an issue, so
> performance isn't the number one priority.

that was clear from the outset.

> Can anyone comment on any of the above or point me in the right
> direction of a good resource talking about the options available (I seem
> to just come across resources that are either just wrong, or only say
> the 100% right way to do it with no flexibility for cost saving - I
> can't afford a 10+ physical disk solution).
>
> Are there any generic examples of how to partition for Oracle in a Linux
> environment?

SuSE supplies documentation here:
http://www.suse.de/en/business/certifications/certified_software/oracle/documents.html

For an Oracle/RedHat install, this is an excellent site: http://www.puschitz.com/

For storage configuration docs, I'd start at http://www.baarf.com.

The OFA standard is a good start, but each version of Oracle supports it differently. You didn't mention what version you were intending to deploy.

start at http://otn.oracle.com and read the concepts guide for the version of Oracle (database server) that you intend to deploy. 9i R2 would be a good choice.
http://otn.oracle.com/documentation/oracle9i.html

btw - have you considered allocating enough space for transaction (archive) logs, physical, logical and rman backup sets on disk? how many oracle homes will be supported (e.g. 8.1.7, 9.2.x, 10.1)?

> Thanks.

hth. good luck. don't be afraid to wipe the box and start over if you're not happy with the results.

Pd Received on Wed Jun 02 2004 - 00:33:33 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US