SSDs - State of the Product?

sdbardwick said:
Bought one from Fry's on a lark. No time to play with it tonight (election party), but here is a HD Tach graph:
Click to expand...

Original post link
My Patriot 32GB SSD that I bought 11/2008 probably died this week. I really cannot complain - it ran 64 bit Vista 24/7 for around 2.5 years without incident. I did move the temp and swap files off of it (to a HDD) early on, so it really didn't get slammed. I thought it would die much earlier; I set up a 30GB partition on the HDD in anticipation back when I installed the SSD. I reinstalled Vista to the HDD partition, and the decrease in responsiveness is immediately noted. Time to break out that spare Intel/Kingston 40V...

(Edit: 'Probably died' because I haven't tested it yet; just yanked it after Vista wouldn't boot; gave error that essential driver file missing/corrupted.)
 
Page file on SSD? If so, how big? Raid stops trim?

Hi
Couple questions.

If you get a RMA'd OCZ Vertex Turbo, that you now have little use for,
and, you wanted to set it up as a page file drive, how big would you make the page file to avoid writing to the same area over and over?

Second:
Is it true that putting SSD's in Raid 0 disables the trim feature, in 7, using Intel X-25M's, or OCZ vertex turbos?

Thanks

GS
 
1. It doesn't matter: the drive's wear levelling dictates which physical cells are actually written. It creates the illusion that you are writing to the same place again and again.

2. Yes, Windows cannot pass TRIM commands to physical drives in a RAID set. That's because Windows sees a single logical drive; it doesn't know which drives held which blocks of a particular file that's just been deleted.

I've read that TRIM won't work even if the drive is dropped from an array, because it's then an array of one, but no idea if that's true or not. Despite Intel's initial claims, apparently the circumstances under which it does work are very narrow indeed (see Wikipedia).

Ddrueding questioned whether Windows would deal differently with a soft RAID, i.e. one created by Windows itself. I have been unable to find any information to confirm this either way.
 
OK:
So a pagefile of let's say 1 gig, on a 30 gig blank drive, would be distributed over the drive, as the wear indicates, not just in a 1 gig space.

The problem I had with the OCZ drive was once it was pulled from the array, and put in the beast, it showed up as 128 gigs, even though it was 30 gigs, and, there was no way to reformat, or allocate to the proper size.

Suggestions?
 
Yes, there are a couple of OCZ utilities. SD linked them somewhere in the mongo SSD thread.
 
drive showed up as dead on my controller card, where it had been fine before.
Better to RMA the drive, or so far, maybe not;-(

Anyway, the replacement drive wasn't formatted, so should be zeroed when formatted, if a long format is used.?
 
I received a 160GB X25-M from Newegg yesterday that was in its approved packaging yet contained a functional install of Vista Home Premium.
 
Mercutio said:
I received a 160GB X25-M from Newegg yesterday that was in its approved packaging yet contained a functional install of Vista Home Premium.
Click to expand...

So those Newegg crooks sold you a used drive as new? Big surprise. :tdown:
 
That's really quite naughty of them (Tech ARP) - there are countless gullible people who will link to that article or make claims on the strength of it.

The author's name is the only really obvious giveaway that I can see.
 
Bozo said:
If it was in "approved packaging" then it came from Intel that way.
Click to expand...

Approved may mean something different than original and unopened. It is more likely that the ripoff occurred somewhere in the supply chain rather than in China (or wherever the drives are manufactured) or the packaging is not genuine. Do we know which container integrity and anti-counterfeiting techniques are used in the packaging?
 
There was an article I read a few weeks ago; boffins at some univ proclaimed that SSDs cannot really be wiped. There are also persistent rumors that cell phones cannot be wiped fully by the user using the factory reset code, and that a shop with the right cable & sw can retrieve all your contacts, emails, and stored info.

I don't know what is true and what isn't, but the truth is out there somewhere.
 
The newer drives i read about are internally encrypted as a default and a clean wipe simply involves throwing away the key. In this case reading the chips would be useless.

On the phone thing i think apple just started using ASLR for user data. Could be wrong about that but essentially also makes illegitimate data retrevial a non starter. I would expect others are doing similar things.
 
Turns out, I know a guy on another forum who actually does work on Intel SSD firmware. Given the exchanges I've had with him in the past year, I have no reason to think that he's lying or joking about it.

He's very down on Intel MLC products. Very down. He's suggested without confirming that they do a lot of potentially unsafe things to maintain their level of performance, and that he knows some things about the lifespan of those products that run counter to Intel's published specs.
 
timwhit said:
Can you reveal what forum this is?
Click to expand...

TotalFark.

The guy I'm talking to is a BSD contributor and former EMC employee. Our conversations to date have suggest that he's done meaningful firmware work in the past on pretty serious storage products and kicked the tires on open source filesystem implementations. I wouldn't be surprised at all to know that he's working on SSDs now. He did say he's under an NDA for the work he's doing now. He's not been very specific about his concerns but he's brought them up often enough that I thought I'd mention it here.
 
timwhit said:
Would you trust an Intel SSD more or less than a Samsung mechanical disk?
Click to expand...

I don't actually have any SSDs doing anything important, now that I think about it. They're boot drives that connect to my local storage configuration, which is triple-redundant and spread out over dozens of drives, plus I have tapes.
 
The more I read the more I appreciate my old X25-E. Intel got that one right from the get go and the write cycles on 50nm SLC must be huge.
 
You must log in or register to reply here.
Back
Top