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: RAID and Oracle Databases

Re: RAID and Oracle Databases

From: Mark Day <mday_at_ucla.edu>
Date: 1998/10/24
Message-ID: <363258ef.550678@news.ucla.edu>#1/1

When he says he will have 5 drives, using Raid 1 and with one hot swap, the hot swap drive just sits and does nothing until a drive in the RAID array dies. Then if the controller supports it, it can automatically rebuild the lost drive from the "Hot" spare. This is usually handled automatically (depending on controller) without the server going down, although there can be serious performance degradation (so maybe you really would have preferred to chance it and wait until night to rebuild the drive). The full redundancy of the RAID array is restored, and you are protected again if another drive fails (as long as the second failure doesn't happen before the automatic rebuild is complete).

This is more cost effective in larger arrays, or groups of arrays. I saw a setup a long time ago which held about 13 or 14 drives, with two RAID 5 sets and a RAID 1 set, and the hot spare could jump in for a dead drive in any of the 3 arrays.

On Fri, 23 Oct 1998 17:07:25 GMT, satar_at_my-dejanews.com wrote:

>Last week, I went from an all RAID5 to partial Raid5 and Raid1. This was
>acomplished by copying the data to a temporary storage device, then
>reconfiguiring the RAID, then restoring the data back onto the RAID. After
>this was done, I had to rename the datafiles to place them on thier
>corresponding RAID drives. I'm not familiar of the logic behind having 5
>drives on RAID1, I didn't know it was possible. You can always use the 4
>drives in RAID1 and if possible, the extra drive as a "dump" drive to store
>your backups (you should keep a copy on tape as well) and other stuff.
>
>Anyway...Good luck to you and to battle the Budget Gods, just have the
>database crash at low peak times, and lie to them it was the lack of drive
>space that caused the crash. (By the way, this is not a good idea and should
>not be taken seriously)
>
>Satar Naghshineh
>In article <WaSX1.971$KM4.625152_at_198.235.216.4>,
> "mikel" <me_at_here.com> wrote:
>> Thanks for the advice.
>>
>> FYI, because I'm not sure if I can squeeze another drive into the picture
>> this year I thought that I would try to get another drive next year
>> (depending on the budget gods of course) and then have the six HD on RAID 1
>> in 1999. I don't know for sure, but I'm guessing it will easier to go from
>> 5 HDs in RAID 1 to 6 HDs in RAID 1 than to change 5 HDs in RAID 1 and in
>> RAID 5 to 6 HDs in RAID 1. Should be just a matter of adding the drive and
>> telling the controller it's there.
>>
>> BTW, as a new comer to the Oracle world, I'd like the thank all the people
>> that contribute to postings by us 'newbies'. This really is a great
>> learning tool (and it's free).
>>
>> Thanks again
>>
>> Mikel
>>
>> 9 usable GB on RAID 1 -> redo log files and index tablespaces (data files)
>> 18 usable GB on RAID 5 ->data files
>>
>Oracle DBA/UNIX System Admin
>Advanced Enterprise Solutions
>(949) 756-0588
>Oracle Re-Seller
>
>-----------== Posted via Deja News, The Discussion Network ==----------
>http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Received on Sat Oct 24 1998 - 00:00:00 CDT

Original text of this message

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