Raid 0 (Stripe) for OS X boot disk? Best Performance and block size

Hi Dan, I have 4 x WD's 1TB's in this baby as well. Firstly for my workflow reasons I am not a fan of using complex RAID systems (raid 1,3,5,6+, and the other b/s configs) as they are wasteful of space and not necessary unless you need instant availability (such as a bank or business). These RAIDx configurations are wonderful if you have spent a lot of money on a very good RAID controller card and a RANK that has hot spares and is more than just a few DDms.... that is excellent at reconstruction in the background that wont impact your work... these are not cheap! (end of rant : ) . I'm more into the creative/post prod field. So:

In my brand new nehalem MAC PRO I use this SOFTWRAE RAID for internal disk configuration. I *have no RAID CARD for the internal disks* - only for EXTERNAL SAS disks in a JBOD enclosure..:
• bay 1 = single JBOD startup disk (startup macosx)
• bays-2-4 = raid0 (raid zero) strip across the remaining 3 DDMs. ( startup OSX as well [IPL?]). this is used for a scatch area for LARGE objects such as VIDE footage and complex objects that fall into the former category. *BLOCKSIZE = 256KB*. The reason is that there are technicall less I/O's to fetch/write the same amount of data where the content is very LARGE.

Simply using such a large block size for typical I.T. based file is wasteful of space, especially for text and small objects.

You will also see a reasonable speed (read & write) in excess of 200MB/sec wi the *AJA system tester* to that stripe across bays2-3 (raid0). (refer http://www.aja.com/html/supportkona3swd.html). +Bare in mind that this I/O rate is not practical in normal work as the AJA driver bypasses the I/I filesystem cache+. It is used for capturing. In any case it is a good indicator of the rates you can get with these WD DDMs and this strip an your new MAC PRO 16 x virt/core Nehalem beast (like mine).

additionally unless you have another equiva;ent fast storage deive inboard or outboard, the speed thing is not really something to be concerned about.

If you have only storage systems (data tape, disk or other) the transfer data rate is simply the lowest common denominator of those devices concerned in the transfer/access. (not rocket science 😉 ). An example would be hooking up an elcheapo usb2 or fw800 storgae disk and performing a transfer. Both of these have lousy data rates well under 32MB/sec (1.5mins/GB or less). As always these things vary to taste/environs/workflows/type of DDMs, rotation speeds and the other 100 parameters... Using a cheap eSATA PCIE HBA and some external SATA disks will see data rates jump to nearly 90MB/sec (no raid) which will be helpful in usong your new found access.transfer speed of your software raid.


I would also propose you use TIME MACHINE to take care of the stuff on the main startup disk (in my case the DDM and file system in BAY1). I haev implemented an Ultrium LTO4 data tape archive system for the stuff on the sowfater raid described above and for the other two disk arrays hanging from a SAS HBA. IF you dont want to go to that trouble, you will need to work out a strategy for looking after the stuff on your sowaftre raid stripe so that your timemachine disk does get full (as I mine used too always). THe latter is a nightmare to resolve.

An option for this is to selctively +EXCLUDE stuf+f you can recreate or get elsewhere from the timemachine backup to limit the items you need to backup (system files and apps and other stuff).

hth

w

HI Dan, yes I DO have the 3 x DDM stripe as a STARTUP ("bootable") OSX 10.5.6 system. I dont have a specific "boot" (or shoe or socks either) drive. THe file systems contain a startup OSX system. This I have two on this new MAC PRO. ( 1 x DDM and 3 x DDMs [the OSX software RAID 0 stripe])

Works fine. 🙂

VEry easy to set up (less than 10 seconds).

blocksizes: for matrail that are large objects, ten use a larg eblock size. If you are using the storaeg system for tiny I.T. type files then use a smaller block size.

the trade off more I/O activity vs possible wasted space.


w

Dan, what kind of performance are you looking for? fast trivials or fast content chunking?

Its always some compromise. I have all my non production/creative stuff (ah la admin + .apps) in the DDM in BAY 1 (a single JBOD file system). I put the other stuff on the 3 x DDM RAI0 software strip and also an external disk array (see recent post).

So I divide the work essence on these specific file systems where I can.

Space is not an issue for me because I use data tape LTO4 ultrium tape drives to archive and recall as I need.

For your performance, just set up and stripe and play with its blocksizes awith various scenarios before loading it up with content. 🙂 USe your .apps and soem tools to try out your situation to see if it solves your concerns.

I think if you go with any kind of redundancy you will lose loads of capacity vs some availability... as always your decision should be based around your workflow and business.

hth

🙂

w

Dear Warwick... I have noticed your replies and take keen interest as I am also contemplating the same raid 0 set up with 4 1TB's. You brought to mind something I did not take into consideration, Time Machine. I really like the simplicity of TM as it saved me once before. So, could you tell me, for photo files, some video, how much does the striping (% wise) improve the accessing and filing of such files compared to no striping but, using internal drives (7200/WD/1TB/Caviar)? I have not done striping before and want to weigh in because of the back up storage issues now. Thanks.

Be aware that striping will increase your exposure to data loss. Striping across 3x HDDs will increase your exposure to disk failure (data loss) by a factor of 3. If you must strip be sure to backup the data on the striped devices. Using Time Machine is a good option but be aware the first TM backup of 3x 1TB HDDs assuming they are moderately loaded up with data will require a large backup device and a fair amount of time to backup initially.

Just a few words of caution..... Good luck. 🙂

Don't use a RAID for everything. Create one RAID just for video files and editing. Boot drives only rarely benefit from changing from default or from RAID with suggested sizes for audio, video, workstation, etc.

Hi D# Shooter, regarding your question,

D3 Shooter wrote:
You brought to mind something I did not take into consideration, Time Machine. I really like the simplicity of TM as it saved me once before. So, could you tell me, for photo files, some video, how much does the striping (% wise) improve the accessing and filing of such files compared to no striping but, using internal drives (7200/WD/1TB/Caviar)? I have not done striping before and want to weigh in because of the back up storage issues now. Thanks.

J_ust give it a try and see if it is worth it for you_. 🙂

Striping:
• just enhances (reduces the access/transfer) because in practice the access is distributed in parallel across several DDM's (Old school but it works great!). I think for video and file work the advantage is that you can access the whole object sooner (rather than faster).
• this distribuition also reduces a load of old style queing on the device ove rthe path. THis was resolved in the late 1980's so no reall rocket science here.

the issues with striping are few and basically over all the raid implementations (except JBOD which of course is not raid) when compared to a single spindle. The discussions are enormous and plentiful via google and experiences and opinions vary widely.

Fir the I.T. peole its the advantage they get for access using a smart disk controller that caches goosies like indexes and stuff so that they can sustain a zillion trivial transactions/sec (i.e. banking & internet stuff).. stuff that is of no interest to me 😟

For the creative people and many applications that are BLOB's (like video, film and remote sensing objects) getting use of the objects sooner (not faster) is of prime importance for workflow efficiencies. If you have this need then striping stiff across disks is for you!

TIMEMACHINE.app works fine as it seems fairly agnostic to whats implemented under the disk file system. MY issue with time machine is that I don't want it looking after my production stuff, only to keep an eye on my admin I.T. type stuff such as ~/ and data data files.

As posted on ths thread:
• availability is the major concern with any file system (cloud or raid or other). RAID with parity schemes and double parity schemes (Raid1,,3,5,6) and implementations such as RAID6+ LSF (log structured file) are all wonderful for this business workflows that need it.
• timely access in a workflow is another
• cost benefits are another

However a *great benefit* for me of *consolidating small storage components under one huge file system is that you dont have to COPY any thing around*. THis is marvelous especially when you think you have to move 2TB's of stuff from one place to anther. THis a takes a lot of time with elcheapo didks that dont have fast interfaces such as SATA/SAS of FC for example.

As always and has been addressed by others on this thread (Hatter) if you lose a component storage device the whole file system is hosed or severely degraded unless you spend a lot of money on full ranks of DDMs with hot spares and a very good RAID controller card. Again its money.

YEah sure you can carry some PARITY RAID implementation around across 3 didks but the storage capacity usage is dreadful. THis is why more complex RAID implemntatiosn are in groups of 10+ dDMs.. (yep poepl can argue.. but this is the mainstream).

My external disk arrays are merely two LUNs (SAS DOMAINSA) that have two file systems implemented using 2 x 4TB 1TBs DDMS - all RAID0 - no parity (no availability) - I just want speed. I look after my own "availability" withm= my archive solution. If the operation dies, I stat again. I'm happy wi that. RAID 5 has write penalty performace hits (well known +update in place+), , RAID 6+ is lousy for huge objects but good for I.T. but ok if you lose two disks in a stripe (RANK).

They all have their flaws... and mirroring a RAID0 (RAID1/0) seems to be popular with storage vendors because they can see you more disk and thats proper business workflow depends on it.

However you can achieve this stuff if you change your workflow slightly.

Other than these the rest is tech specs and stuff under the cover.

So you what is right for you and your business.

I dont like spending money on nasty elcheapo FW800 LeCIE disk enclosures with the their junky components and their ilk having been done badly on several corrupted devices and lsing TB;s of content - this is why I invested in a high speed LTO4 ULtrium data tape archive solution.

sorry for long post..

w

Warwick, thanks for the details. I agree that tape is the ultimate backup and archive. But, I feel comfortable with storage of my files, photos, video on multiple HD in various locations. I am interested in raid 0 for access during editing photos and sometime doing video. Which leads me to ask you, in your experience.... +how much sooner (percentage) does striping give you over accessing conventional file storage or writing? Is it worth the trouble? And... if bay 2 and 3 are for striping, could bay 4 be for raid 1 as backup? And, one other, curious which processor you think is right for photo editing and some video, 2.26 (8core) or same as your 2.93?+ Thanks.

Message was edited by: D3 Shooter

That's what I am concern about, 2 more drives operating sequentially, and what would happen if one fails, does the other write all files until you can get the 2nd drive in raid format up again? Talk about back up, where do the files go then until the HD is replaced?

D3 Shooter wrote:
Warwick, thanks for the details. I agree that tape is the ultimate backup and archive. But, I feel comfortable with storage of my files, photos, video on multiple HD in various locations. I am interested in raid 0 for access during editing photos and sometime doing video. Which leads me to ask you, in your experience.... +how much sooner (percentage) does striping give you over accessing conventional file storage or writing? Is it worth the trouble? And... if bay 2 and 3 are for striping, could bay 4 be for raid 1 as backup? And, one other, curious which processor you think is right for photo editing and some video, 2.26 (8core) or same as your 2.93?+ Thanks.

Message was edited by: D3 Shooter

I'd have to say if you are worried about losing spindles (DDMS) in any kind of stripe then simply don't do it.

THese days I see that cheap disks are less reliable the more capacity they seem to have to keep the prices down.

THe big advantage if having these disks as +single logical file system+ is that you don't have to COPY objects around to manage the storage space.

In your case I would suggest you just try it out for a while. You can always revert back to 3 single disks that give you three file systems. You could use any of these as a TIMEMACHINE backup or you could ALSO consider a MIRROREd internal disk set up (disk utility.app again).. say BAY3 and BAY4 are mirror. YOu reduce your storage capacity by 25% but that's how it goes. You are only protecting yourself against a faulty disk (wise these days). but not faulty data.. hence TIMEMACHINE.app is great!


*MACPRO for photo editing*. More than enough.. but I'm not into it full time for photo editing.. ask others optinoin... I think the lower end model is more than you need. You might want to consider the ATI 4870 graphics card over the other one. Think also of a dual screen set up.

I've not bothered to look at the NVIDIA cards.

The ATI 4870- has a DVI and an  Apple mini AV port on it. You could utilise two display screens for your photo applications (LM, APerture etc, PSHOP etc etc).

Like many I bought this 2.93Ghz model with the dual quad core t for the specific purposes of transcoding and rendering (SHAKE etc) and effects work where I dont want to wait too long to get a shot done or a package distribution performed..

If you do a limited amount of basic video editing right up to 4K you'll be ok with the 2.26. I edit this stuff also on a MACBOOK PRO.. its fine too 😉 .

HTH

w

A. Your system only needs to be backed up once, then updated anytime before you install updates. And keep everything else off your boot drive.

B. Torture your drive (zero it out or more) before you format and install anything. And run TechTool Pro or something to do a full surface media scan.

C. Four disk drives each doing its own task, "four wheel drive" when done right.

D. While louder, the WD 10K drives seem to be made to higher quality, but too small for media maybe, great for system drive, for scratch.

E. Stripped RAID was often used just for scratch. And to expand beyond the capacity of a single drive. Backed up before work, erased if scratch between projects or steps.

F. Start with just a scratch stripe array for experimenting.

G. Add a 2nd stripped array volume, another two drives, and now copy between the two, notice how much smoother and faster, and how the sytem drive that isn't part of your array, is still responsive and free.

H. Large 4GB files can and do otherwise cause a system to 'come to its knees' as the system isn't fully 64-bit, and while the Finder was copying, it had trouble doing anything else. Another place where an array can ease the burden some.

I. A mirror is not a backup. And in some servers, you use 3 drives, not two. With a stripe, there is just one volume over n-number of drives and the whole array needs to be rebuilt.

J. Drive failures happen, but unless you want total 100% real-time backup, restoring a drive to where you were an hour or a day ago should just be a matter of swapping / formatting / restoring and be back on your way in an hour.

K. Always have a minimum of two backup sets for any drive or volume.
Always have a drive you can "drop in" if needed with a backup on it and then just run an update perhaps.

There are some guides to storage, to optimizing for photoshop, a good place to look and read is the Help topics on MacGurus left side column.

Warwick and Hatter... +Great stuff, I appreciate all that you contributed, a lot to take in and to try... thanks!+

Hi, I have a new MP 2.66 (dual) that is solely for FCS and video in general. I striped (0) two Velociraptors for boot and wares and now I'm wondering about a maintenance schedule. What can be used to maintain the healthy snappy performance?

What utilities can be used to defrag and repair disk permissions? Can I still use DU on two slices without fouling things up? Can iDefrag be used? Other?

Don't mean to hijack this thread, but this info would be relevant here.

BTW, two internal 1TB Barracudas are used for video storage.

Thanks