dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
24
share rss forum feed


koitsu
Premium,MVM
join:2002-07-16
Mountain View, CA
kudos:23

1 edit
reply to mattrixx

Re: TRIM / "garbage collection" with RAID0

GC is separate -- handled by the drive firmware itself. The FTL inside of the SSD has to keep track of every LBA written to.

Without TRIM, the drive has to make decisions itself to determine when an LBA is no longer actually used. I'm sure you can imagine how dangerous/troublesome this can be (for software design).

The more common problem with drive GC is that it suspends any I/O transactions while its doing the clean-up -- this has been shown time and time again in benchmarks where I/O rates suddenly drop to zero for a short time then jump back up again).

Bottom line is that GC isn't as efficient as the OS saying "I'm done with this LBA, you can reclaim it", that way the drive knows for certain when the LBA can be removed from the FTL map (or re-purposed).

Without TRIM, your drive will, over time (greatly depends on the capacity used and the workload), begin to decrease in performance due to the FTL map growing very very large (many LBAs in the list to iterate over). Zeroing the drive doesn't help this either; the only way to clear the FTL map is to issue an ATA SECURE ERASE transaction (this is not the same thing as writing zeros to every LBA), which is what's required when relying purely on GC for long periods of time.

These are concepts that apply universally to SSDs (meaning, regardless if RAID is in use or not).

TL;DR -- TRIM is important, but you can make do without it if need be. Just be aware of the repercussions.

As for the P8Z68-V Pro -- it's driven by an Intel Z68 chipset, which means that as long as you're using recent Intel RST drivers, TRIM support is provided for RAID 0 arrays. So rather than worry about any of the above, just make sure you're using recent RST drivers and don't worry about the rest.
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.


mattrixx

join:2004-02-18
Orland Park, IL

Thanks Koitsu



Octavean
Premium,MVM
join:2001-03-31
New York, NY
kudos:1

2 edits
reply to koitsu

said by koitsu:

As for the P8Z68-V Pro -- it's driven by an Intel Z68 chipset, which means that as long as you're using recent Intel RST drivers, TRIM support is provided for RAID 0 arrays. So rather than worry about any of the above, just make sure you're using recent RST drivers and don't worry about the rest.

I was of the understanding that only 7 series chipsets such as the Z77 currently had TRIM support for RAID 0 configurations with X79 chipset support soon to come.

quote:
The requirements for RAID-0 TRIM support are as follows:

A 7-series motherboard (6-series chipsets are unfortunately not supported).
Intel's Rapid Storage Technology (RST) for RAID driver version 11.0 or greater (11.2 is the current release)
Windows 7 (Windows 8 support is forthcoming)

The lack of support for 6-series chipsets sounds a lot like a forced feature upgrade. Internally Intel likely justifies it by not wanting to validate on older hardware, but I don't see a reason why TRIM on RAID-0 wouldn't work on 6-series chipsets.


Therefore the Z68 chipset wouldn't be supported.

***edit***

»TRIM on RAID 0 SSD Array Finally

***edit again***

If I were to guess, I would say that the Intel RSTe (Enterprise) update for the X79 / C606 chipset will bring TRIM support for RAID 0 configurations at or near Windows 8 release on October 26th.

That's just an off the wall guess though.


koitsu
Premium,MVM
join:2002-07-16
Mountain View, CA
kudos:23

Thanks for educating me, Octavean See Profile.

I didn't realise the chipset series was so important. In fact, I'm not quite sure why it would be -- think about what actually drives the RAID bits of the chipset when it's in RAID mode: the option ROM (think: "RAID BIOS") is what handles it. That's quite simply just a bunch of x86 code.

It sounds to me like Intel doesn't want to deal with implementing the necessary option ROM bits on older chipsets. They would have to go to every motherboard vendor and tell them to update their system BIOSes (which is what contain the option ROM code for RAID as well as the option ROM code for AHCI), etc. etc... I'm willing to bet that starting with the 7-series chipsets, from the get-go, the option ROMs they started with (thus distributed to all manufacturers) contained the necessary code for this to work (even though the driver didn't). I bet Intel wants to "start with a clean slate". Or it could be that Intel just wants people to buy new crap. Or both. Sigh.

Basically what I'm saying is that there isn't anything silicon-level that would inhibit this from being done on 6-series chipsets or earlier. *grumbles* And without a SATA-capable ATA protocol analyser, there's no actual way to tell if their "7-series and newer only" claim is even legitimate.
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.



norwegian
Premium
join:2005-02-15
Outback

Not that I have answers for the OP, but what is engineering possible doesn't always equate to sales koitsu See Profile.

Lovely world we are in and all that hype.

From that though, on board raid is not a dedicated raid card - would that make any difference for the OP if he/she was specifically looking for TRIM?
--
The only thing necessary for the triumph of evil is for good men to do nothing - Edmund Burke



Octavean
Premium,MVM
join:2001-03-31
New York, NY
kudos:1

said by norwegian:

From that though, on board raid is not a dedicated raid card - would that make any difference for the OP if he/she was specifically looking for TRIM?

Unlikely since a RAID card would not have TRIM support in a RAID 0 configuration regardless of the chipset whereas select Intel 7 series chipsets do.

mattrixx

join:2004-02-18
Orland Park, IL

Well thanks for your most informative links and replies Octavean.

Looks like I have opened a can of worms here regarding RAID0 and TRIM.

At the time of my RAID creation and Windows installation, I was led to believe RAID0 and TRIM together was possible, and had NO inherent problems!

Can you further enlighten me with what kind of "issues" I can expect to encounter with my present Win7 installation on to (2) 64GB Crucial M4 SSDs in a RAID0 array?

Thanks again



Octavean
Premium,MVM
join:2001-03-31
New York, NY
kudos:1

Unfortunately I am not entirely sure what you might encounter over time.

I have a number of different SSD models of which a Crucial M4 256GB SSD is one of them. However, I don’t currently have an SSD RAID 0 array setup. Although I have run such configurations in the past and have had no real problems or major concerns over performance loss:

»SSD Raid 0 Array 4x60GB Setup Test Run

quote:
Author: Allyn Malventano
Date: February 17, 2010
Subject: Storage
Manufacturer: Kingston

TRIM

I don't have and pretty charts or graphs to explain this next part, but I will share an observation I made during my fragmentation testing. When running my fragmentation tool, I observe IOPS drop as the drive becomes more and more overloaded with the task of tracking the random writes taking place. Here the JMicron controller behaved like all other drives, but where it differed is what happened after the test was stopped. While most other drives will stick at the lower IOPS value until either sequentially written, TRIMmed, or Secure Erased, the JMicron controller would take the soonest available idle time to quickly and aggressively perform internal garbage collection. I could stop my tool, give the drive a minute or so to catch its breath. Upon restarting the tool, this drive would start right back up at it's pre-fragmented IOPS value.

Because of this super-fast IOPS restoring action, and along with the negligible drop in sequential transfer speeds from a 'clean' to 'dirty' drive, it was impossible to evaluate if this drive properly implemented ATA TRIM. Don't take this as a bad thing, as any drive that can bring itself back to full speed without TRIM is fine by me, even if that 'full speed performance' is not the greatest.

This type of self-healing (i.e. without needing TRIM) is great for those wanting to run a few SSD's behind a RAID, since no RAID implementation is currently capable of passing TRIM from the OS to the arrayed SSD's. Better yet, considering this drive is tailored to the budget crowd who may very well still be running XP or Vista, it's good to have a few choices that don't require TRIM to maintain decent levels or performance.


»www.pcper.com/reviews/Storage/Ki···rance/Fr

This was applicable for the SSD model that I was chiefly using in the RAID 0 array which was the Kingston SSDNow V Series SNV-S2 64GB (3x but not applicable to the single 60GB OCZ Agility in the array).

Personally I would just keep an eye on performance overall. That does not mean benchmark the configuration aggressively. It just means take note via empirical observation of your system in everyday use and maybe one or two benchmarks a month (or every other month) to look for a delta. Some SSD models (controllers) may degrade performance somewhat as the storage is fills up (independent of TRIM and Garbage Collection concerns) so keep that into consideration as you record your data.

As SSD prices fall you may find that you can replace your 2x64GB RAID 0 array relatively cheaply. I replaced my 4x60GB RAID 0 configuration with a single 240GB SSD that was about the price of one of the single 60GB / 64GB SSD models when they were new.

I’ve personally been thinking about replacing my Asus P8P67 Pro with a Gigabyte GA-Z77X-UP5 TH (dual Thunderbolt) for ~$239.99 or a GA-Z77X-UP4 TH (dual Thunderbolt) $184.99.



norwegian
Premium
join:2005-02-15
Outback
reply to Octavean

said by Octavean:

said by norwegian:

From that though, on board raid is not a dedicated raid card - would that make any difference for the OP if he/she was specifically looking for TRIM?

Unlikely since a RAID card would not have TRIM support in a RAID 0 configuration regardless of the chipset whereas select Intel 7 series chipsets do.

Thanks.
--
The only thing necessary for the triumph of evil is for good men to do nothing - Edmund Burke



koitsu
Premium,MVM
join:2002-07-16
Mountain View, CA
kudos:23

What Octavean See Profile said is generally correct -- as of this writing I don't know of any hardware RAID controllers that provide TRIM.

All the manufacturers of HBAs (ex. Highpoint, Areca, 3Ware, LSI) say the same thing: we acknowledge the importance of TRIM in a RAID configuration, but driver-wise everything is implemented via SCSI emulation and TRIM is purely part of ATA protocol, thus getting it to work is somewhat tricky.

Solving the problem at the firmware level, or even silicon-level, is ideal -- it just requires that everything be designed a very specific way (from the ground up). It involves tracking LBAs to some degree. (See, this is where I wish I worked at Intel -- I would love to know what all the engineers had to do to get this to work on their recent chipsets + RST. )
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.



norwegian
Premium
join:2005-02-15
Outback

RST:

So would this also be a reason to obtain Intel SSD's too?

Or does it work across differing manufacturer platforms, as long as the chipset is Intel on the motherboard and the SSD isn't too overly relative to the raid?
--
The only thing necessary for the triumph of evil is for good men to do nothing - Edmund Burke



Octavean
Premium,MVM
join:2001-03-31
New York, NY
kudos:1

1 recommendation

As long as the SSD(s) supports TRIM and all the other requisites are satisfied then there should be no problems. The SSD(s) does not have to be of Intel manufacture.