RE: Solid State Drives

From: Freeman, Donald <dofreeman_at_state.pa.us>
Date: Fri, 1 May 2009 11:04:14 -0400
Message-ID: <55264C4C0484A547B34C0B1A28E219EA338616653E_at_ENHBGMBX01.PA.LCL>



OK, I think I understand now. Each of these new drives is equivalent to a spindle and you mirror and stripe exactly the same way.

From: Matthew Zito [mailto:mzito_at_gridapp.com] Sent: Friday, May 01, 2009 10:56 AM
To: Freeman, Donald; Andrew Kerber
Cc: oracle-l_at_freelists.org
Subject: RE: Solid State Drives

No, you'd still buy 2 x drives for mirroring, and 3+ x for RAID5. Remember that the physical interface from the drive to the array is still a SPOF, so you can't depend on the internal isolation within a drive.

Thanks,
Matt



From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Freeman, Donald Sent: Friday, May 01, 2009 10:51 AM
To: 'Andrew Kerber'
Cc: Oracle-L (oracle-l_at_freelists.org) Subject: RE: Solid State Drives

So, if I bought one of these 4TB monsters I'd just split it into 2X2 TB mirrors? Does it have to be striped or that doesn't make sense any more?



From: Andrew Kerber [mailto:andrew.kerber_at_gmail.com] Sent: Friday, May 01, 2009 10:42 AM
To: Freeman, Donald
Subject: Re: Solid State Drives
Mirroring, etc is still required. Like most modern storage, the arrays themselves normally handle this though.

On Fri, May 1, 2009 at 9:38 AM, Freeman, Donald <dofreeman_at_state.pa.us<mailto:dofreeman_at_state.pa.us>> wrote: But what about mirroring? Are these devices so reliable that we won't have to make redundant copies of the data to prevent loss? What's the architecture look like?



From: Andrew Kerber [mailto:andrew.kerber_at_gmail.com<mailto:andrew.kerber_at_gmail.com>] Sent: Friday, May 01, 2009 10:18 AM
To: Freeman, Donald
Cc: Oracle-L (oracle-l_at_freelists.org<mailto:oracle-l_at_freelists.org>) Subject: Re: Solid State Drives
I wrote an article in IOUG Select journal about Solid State. I really think Solid Statie is the coming thing. At this point, it is too expensive to put all your storage on SS, but if you put your popular tables on it, you can see substantial speed improvements.

If they can resolve the write issue, I could see SS really helping to reduce pinging in RAC by making the IO to disk almost as fast as the IO to cache, thus reducing the required cache sizes and the pinging caused by that. Redo log would also be a good usage for SS, reducing the time to switch logs.

Currently, the developers are centered on leveling the writing, that is making the level of writes the same across all portions of the SSD. That is showing promise in extending the life of the SS storage. Solaris is working on a major initiative in SSD, I dont know how that will be affected by their purchase by Oracle On Fri, May 1, 2009 at 9:09 AM, Freeman, Donald <dofreeman_at_state.pa.us<mailto:dofreeman_at_state.pa.us>> wrote: Has anybody given any thought to where we are going as SSD's get cheaper and bigger? We've been going round and round at my shop with discussions about RAID, other disk allocation issues, fights over storage. I mean we seem to spend a lot of time on that issue. I saw that IBM is testing a 4 TB SSD. I was wondering if you'd have to mirror that, What kind of reliability we would be getting. No more RAID discussions? I've heard there is a finite number of times you can write to it. What's the upgrade path here?

--
Andrew W. Kerber

'If at first you dont succeed, dont take up skydiving.'



--
Andrew W. Kerber

'If at first you dont succeed, dont take up skydiving.'

--
http://www.freelists.org/webpage/oracle-l
Received on Fri May 01 2009 - 10:04:14 CDT

Original text of this message