| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: I have a clarification
Hi Steve & list,
Sorry for the delay in posting, the day job tends to take over life quite a bit. My comments to yours below.
Hope it helps,
Gaja
> Why search google when I can just ask you? :-)
>
Flattery will get you nowhere....;-)
> We had some NetApp sales drones here last week and
> they made tremendous
> claims about their performance with Oracle. The
> sales engineer claims to
> have been an Oracle DBA for a few years and he made
> a compelling case but I
> haven't been able to get any performance feedback
> from current DBA's with
> NetApp filers in production. Do you have anything
> you can share as regards
> Oracle I/O performance on NetApp Filers which are a
> RAID4 implementation?
>
Sure, the performance of the NetApp. Filers is pretty good if you configure it right. Pay heed to some of the basics in DBA101 - such as separate DATA from INDX. This can be done with their volume manager and I urge you to do that. Separation of write-intensive components on separate volumes will also help. Creating one huge volume with all the disks and not separating contending components can and will cause disk-level "head thrashing". This will cause I/O bottlenecks regardless of your storage vendor.
> BTW, got your book... great job!! You make mention
> of NetApp and RAID4 in
> your book. You say RAID4 is like RAID3 (except it's
> block level parity) so I
> assume it performs like RAID3 which you state is
> better for DSS but not as
> good for OLTP. NetApp claims to have overcome RAID4
> performance issues with
> their wafl architecture. Any feedback on this?
Thanks for the kind words about the book. Could not have done it without Kirti's support and efforts.
Yes, RAID3/4 primarily caters to applications that perform bulk sequential reads with a few number of users. But, with the WAFL (Write Anywhere File Layout, referring to the act of writing filesystem meta-data blocks anywhere in the file) architecture, the overhead associated with the management of writes to stripes of data within the RAID4 volume is greatly minimized.
Very briefly, the write-anywhere design allows WAFL to operate efficiently by scheduling multiple writes to the same RAID stripe (when applicable) to minimize the write penalty incurred with updating blocks independently and repeatedly in a RAID stripe. This along with the fact that NFS write requests are grouped, scheduled and then written in a meaningful order, makes it a very viable solution even in "write-intensive OLTP environments". However, the optimal configuration of the network segment on which the Filer is put, is quintessential to optimal performance. Follow NetApp's recommendations for that and you should be fine.
>
> Waiting for your next book,
> Steve Orr
>
Oh don't go there now, I am just getting back my life
after 6 months...;-)
Cheers,
Gaja
>
> -----Original Message-----
> From: Gaja Krishna Vaidyanatha
> [mailto:oraperfman_at_yahoo.com]
> Sent: Wednesday, July 11, 2001 7:22 AM
> To: Multiple recipients of list ORACLE-L
> Subject: Re: I have a clarification
>
>
> Peter & list,
>
> NFS 2.0 and up is TCP-based and hence has the
> required
> reliability to support "acks and error checking" of
> network I/O calls. Having said that the only vendor
> that I am aware of where NFS-mounted filesystems is
> "certified by Oracle" for an Oracle database is
> Network Appliance. In the case of NetApp. because it
> is a NAS (Network Attached Storage) device and the
> filesystem is maintained and supported on the Filer
> rather than on the "host". In a very simplistic
> sense,
> the host has a pointer to the NAS device on which
> the
> filesystem is maintained. For more information on
> NAS
> or anything else I urge you to search on
> www.google.com.
>
> And yes, you can hookup the NetApp. Filer to a Sun
> Solaris box for sure. Don't know about Linux, but
> you
> can check NetApp's website to verify that.
>
> Cheers,
>
> Gaja
-- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: Gaja Krishna Vaidyanatha INET: oraperfman_at_yahoo.com Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051 San Diego, California -- Public Internet access / Mailing Lists -------------------------------------------------------------------- To REMOVE yourself from this mailing list, send an E-Mail message to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).Received on Thu Jul 12 2001 - 18:38:31 CDT
 |
 |