Slow Raid

Our Raid has become extremely slow recently. Testing on another Raid shows that operations that should take less than half a minute are taking 11 minutes. I have tried various cache settings and reconditioning the Raid with no improvement. I am thinking a drive is behaving poorly, but I wonder how you could identify the culprit drive since I am getting no errors in the logs.

How full is the RAID?
Sounds like it's over the advised 85% full. The behavior you're describing is exactly what you see at over about 92% full.
=Tod

Similar Messages

Slow RAID speeds

I just installed 4 Western Digital WD2500KS SATA 2 drives and RAID'd them using disk utility. Problem is I do no think I am getting the speed that these drives should be putting out.
Using Blackmagic Disk Speed Test, testing one single WD2500KS drive I get
a Read rate of 40.5 MB/s and a write of 50.2 MB/s
I then RAID 4 of the same drives and test with same utility and get
a Read rate of 116.0 MB/s and a write of 179.6 MB/s
I used Apple Disk Utility to RAID the drives and selected striped option and
Mac OS Extended (not journaled)
any idea what this read rate is so slow? should it not be roughly 4 times the single drive rate, somewhere around 160 MB/s?
btw the drives did not come with jumpers and I did not use any, just put them in. I have the boot drive connected to the extra SATA2 port on the motherboard, the 4 drives are in the 4 Mac Pro drive sleds.

These are the Caviar SE16 series. If I was choosing RAID, I would have chosen the WD RE line that are built for and have features to help with RAID.
I wonder if that other SATA port you use for boot drive is throwing a monkey wrench in the works??
Not uncommon for some users to zero a drive before using, to insure that it passes and might help map out any bad block if there are some.
I bought a couple WD RE 160GB drives and they are "okay" but I was expecting more than the 55-60MB/sec (I get 60-62MB/sec from WD RE 320GB model).
I don't have Blackmagic, so I rely on Intech QuickBench and ZoneBench, the later really helps isolate any area of the drive that isn't performing properly and where I/O falls off, as well as an excellent method to pretty much "burn in" a drive with 50 zones and 30x r/w of a 500MB test file thru-out.
Maybe is has to go through nn restarts or something and operates in some safe mode initially? Can you make sure it doesn't have a jumper?
Mac Pro 2GHz 2GB WD Raptor/Caviar RE 320 Mac OS X (10.4.7) 2 x APC RS1500s

Extremely slow RAID reconstruction on px2-300d

Hi.
My rather new px2-300d suffered some weird lock few days ago (dec 20th): the front display was locked to a specific hour/time, alert led flashing, network not responding at all.
So few days later (dec 23rd) I could physically access the device I forced shutdown and restarted it, RAID was inconsistent and it started autorecovery of the 3TB disks. Well, the recovery is STILL RUNNING NOW, at ~65% progress.
This doesn't look normal to me, the crash at first and then this extremely slow recovery, I've never seen such slowness even on very very old hardware.
Also, the NAS itself is very slow when syncing my 400GB data with rsync over nfs.
The firmware is not really up to date, I planned the upgrade as soon as the rebuild is complete.
Is there anything I can check to see what's going wrong? or is it just a problem with the disks?
thanks
Solved!
Go to Solution.

Hello maxxer
With any of the LenovoEMC/Iomega lifeline devices, the firmware is loaded to the hard disks. Even the diskless models load the firmware to the hard disks on the initial setup.
Because of this in some situations a disk error could cause the firmware to behave irratically. With the unit booting from the disks, a drive issue could also affect data transfer and overall performance.
If you have been trying to copy data over while a reconstruction or data verification is in progress, you will see a significant performance hit until it completes.
I recommend contacting support when you are able to. Usually screen shots from parts of the web ui are gathered in addition to the event log and possibly a dump report file from the unit to further diagnose what is happening.
LenovoEMC Contact Information is region specific. Please select the correct link then access the Contact Us at the top right:
US and Canada: https://lenovo-na-en.custhelp.com/
Latin America and Mexico: https://lenovo-la-es.custhelp.com/
EU: https://lenovo-eu-en.custhelp.com/
India/Asia Pacific: https://lenovo-ap-en.custhelp.com/
http://support.lenovoemc.com/

Slow RAID after 10.4.3 update

Powermac G5 with PCI-X Fibre Channel card connected to RAID with optical cabling. Before the 10.4.3 update the connection was fast. After the update the download speed is now 5-6 times faster than upload speeds. Tried to reinstall the driver from the CD, but it will not install. Besides the useless Fibre Channel utility are there any other testing utilities for OS X? Is there any way to reinstall the driver without reverting to 10.4?

Hi,
I have the same problem. My iMac also hangs for about one minute at the same place;
"Nov 23 16:07:04 localhost kernel[0]: IPv6 packet filtering initialized, default to accept, logging disabled
Nov 23 16:07:04 localhost kernel[0]: UniNEnet: Ethernet address 00:11:24:86:33:46"
but with a different MAC address ofcourse.
I 've run the maintenance scripts and repaired permissions using Onyx but it did no difference.
Any other ideas?
iMac DV/450

Suggested methods for full backup of XServe RAID data

I know this is only peripherally related to the discussion topic, but since every other suggestion posted here is followed by the disclaimer that you should make a full backup of your data before proceeding with any major operations on your RAID arrays, I'd like to know what more experienced admins do in order to create a full backup for reasonably fast recovery in case of substantial data loss during maintenance/repair.
Our current "backup" availability is incremental optical disc archival (our data is mostly "write-once"), but this isn't entirely practical for recovery since it's over a terabyte of data. Since the connected server has a free hot-swappable SCSI drive bay as well as an interface for external SCSI devices, not to mention the fiber channel and ethernet interfaces, the options that I'd consider in order would be:
1. A handful of 150-500 Gb SCSI hard drives, rotated out of the hot-swappable bay
2. An external tape drive attached to the SCI interface (with appropriate tape size, maybe the LTO-2 with 200Gb native capacity?)
3. Some other external SCSI storage device
4. Larger optical disc archival (I hear there are technologies arriving in the near future)
5. Network-based option; remote seems impractical due to sheer size, but perhaps local?
The idea is to make a full backup (long-term solutions are superior of course) of 1-2 Tb of data on the XServe RAID before attempting major surgery. Suggestions for common, accepted, tested, efficient methods for accomplishing this would be greatly appreciated. I apologize if this thread isn't on-topic enough for some of you.
-Brian

Brian,
Tape IMO is kinda yucky (to steal a term from your average 3 year old). It's fairly slow to back up to, it's very slow to restore, and it's actually not that reliable by itself (I worked with a large enterprise customer who said their backups were successful about 70% of the time (!!!)).
That said, tape has the advantage that you can offsite it and archive it very cheaply, and the media are fairly cheap, so you can make lots of backups, so if one fails, you probably can restore the data from another tape.
Disks are more expensive initially, but end up being pretty reliable, and you get a lot more flexibility (plus, they're fast).
An emerging "best of both worlds" backup strategy is what's called disk to disk to tape, where you typically back up to another large "disk," for example a second Xserve RAID. Data is then backed up from the second disk to tape, which is taken offsite... thus tape is used for what it's best at (offsite archival). Restore can be from disk in most cases, which is 10-20x faster than restoring from tape. People use software packages like Netvault's Bakbone or Atempo's Time Navigator, which can handle the whole process, and it works quite well. The backup disks (e.g. the RAID) can be onsite, or can be at a backup site a couple KM away, attached via optical (this is preferable, for DR reasons).
For cases where a second Xserve RAID is prohibitively expensive, cheaper (and slower) RAID 5 enclosures like Wiebetech's RAIDtech can provide a large (say, 1.6 TB) RAID 5 volume, accessible over FW800 or SATA (not sure if they have a SATA-based one yet).

RAID card or Disk Utility RAID - Need simple comparison please.

Folks, I am trying to make sense of the necessity or lack thereof for Apple's RAID card. Can someone provide a simple breakdown or comparison of the RAID card and/or loading the Mac Pro with drives and striping them using Disk Utility. I was under the impression that I could buy a Mac Pro, fill up the drive slots and just have Disk Utility set up the RAID. I'll be working with HD.
Enlighten me please

Can someone provide a simple breakdown or comparison of the RAID card and/or loading the Mac Pro with drives and striping them using Disk Utility.
Users that are working with uncompressed HD 1920x1080 10bit RGB video usually aim for storage speeds of 240MB/sec or more. Obtaining this level of performance across the RAID is the trick.
The Apple RAID 5 card can provide 306MB/sec. when configured as a striped RAID set, using four Seagate 250GB model 7200.10 internal hard drives with 16MB cache, when the volume is empty. By the time the striped RAID set reaches 80% full performance drops to 214MB/sec. A setup that can provide 240MB/sec. when the volume is 100% full will provide a more reliable configuration for 1080 uncompressed HD video processing.
Disk Utility can provide this same level of performance without the RAID 5 card.
So what is the advantage of the Apple RAID 5 card?
The redundancy of RAID 5 can add a layer of protection against the failure of a single hard drive. RAID 5 can be rebuilt whereas RAID 0 provides zero data protection. The problem with RAID 5 in a four drive setup is that one disk is required for storing parity data. This leaves the RAID 5 slower as only 3 disks can be used for performance. In a RAID 5 four drive setup the top speed available will usually be approximately 210MB/sec. and when full closer to 180MB/sec. As you can see RAID 5 adds some data protection but the price is lower performance. That is the feature that the Apple RAID 5 card offers combined with a bootable internal solution.
Do I need 240MB/sec. performance?
Users that are not working with uncompressed HD 1920x1080 10bit video may find slower RAID performance will work for them. HDV requires approx. 25MB/sec and DVCPRO HD needs 100MB/sec.
On the other hand, when I am working with large video files the faster the RAID, the easier it is to work with large files. So while I may not have dropped frames with slower DV formats I still
prefer to work with as fast of a RAID configuration as I can justify.
Other Options?
The Mac Pro has many superior performance options available for creating fast RAID volumes. My current favorite setup is the eight channel Areca ARC-1221x RAID 6 controller paired with an external Enhance E8-ML enclosure. You can see an AMUG review of it here:
http://www.amug.org/amug-web/html/amug/reviews/articles/areca/1221x/
This setup provides a RAID 6 array with twice the redundancy of RAID 5. It also supports up to eight hard drives which significantly enhances performance. Using eight Seagate 320GB model 7200.10 hard drives in a RAID 6 configuration with the ARC-1221x provides over 430MB/sec. when empty and over 220MB/sec when 100% full. Up to two drives can fail and the RAID can still be rebuilt. The ARC-1221x is available to AMUG members for $680 until the end of the month here:
http://www.tekramonline.com/amugpromos.html
The Enhance E8-ML 8 bay enclosure is $595. Details are here:
http://www.amug.org/amug-web/html/amug/reviews/articles/enhance/e8/
You will also need two external Mini-SAS to Infiniband cable model Ext-MS-1MSB. I got mine here:
http://www.amazon.com/exec/obidos/ASIN/B000FUOMO0/arizomacinusergr
So for $1340 you get an 8 bay RAID 6 setup that provides awesome performance, supports RAID levels 0, 1, 3, 5, 6, 10 or JBOD and can even be setup to boot the Mac Pro with any of these RAID configurations. I think this is a great setup.
More Options
If the user already has SATA PM enclosures and wants to add external RAID 5 capability for minimal cost, HighPoint has introduced the new RocketRAID 2314.
http://www.amug.org/amug-web/html/amug/reviews/articles/highpoint/2314/
The HighPoint RocketRAID 2314 only costs $200 but it adds nice RAID 5 performance with SATA PM enclosures when using the new Mac version 2.11 driver. Using two SATA PM enclosures with 5 hard drives mounted in each enclosure for a total of 10 drives can provide RAID 5 performance of over 370MB/sec when empty and over 318MB/sec when 100% full. It doesn't offer RAID 6 or boot capability but this is very nice RAID 5 setup.
The card costs $180 on sale here:
http://www.amazon.com/exec/obidos/ASIN/B000NAXGIU/arizomacinusergr
Two quiet Sonnet 500P five bay enclosures will run $1000
http://www.amug.org/amug-web/html/amug/reviews/articles/sonnet/500p/
So for $1180 users have a high performance RAID 5 setup that supports up to ten SATA hard drives. The RAID 5 performance is twice as powerful as an internal four bay solution. Plus, this setup allows users to still utilize the Mac Pro internal bays for importing more data or for backup.
The Apple RAID 5 solution is a nice one. It just seems a little pricey compared to other faster solutions that are available. However, if Apple ever came out with a new RAID 6 card that used the internal 4 ports plus offered two external mini-SAS ports for a total of 12 drives, that would be a product totally worthy of a $1000 premium.
Happy hunting!

Creating RAID 0

Here's the setup: two identical 40gb disks, one used as large storage the other used for the OS and 30gb stripe.
I'm having trouble getting the correct file size reported back when I use the df -k command. Instead of getting a 66 gb virtual drive, it only reports the first volume of 36gb. I know they are both attached. Here is what I have done.
1. I've created partitions and added the newfs to each of the partitions.
2. I have set up the metadb,
3. I have run the metainit and received the concat/stripe is set message. I've called it d23 as a drive.
4. I've edited the /etc/vfstab file to include the new drive.
It's just not showing up as a 66 gb hard drive.
If anyone can see a mistake or knows of a better way of doing this, let me know. I'm going bald with frustration...
ps. I've tried using Solstice Disk Suite 4.2.2 and SMC. No luck with them either.

You cannot use an existing drive as part of a RAID without it being erased. So you need to make a backup in order to do what you want. However, making a RAID from an SSD and an SD card will result in a very slow RAID. It is not a good idea. RAIDs should be created with similar if not identical drives for best performance. You would be better off replacing your SSD with a larger one if that is possible in your model. Or, just get an external hard drive, forget the RAID and the SD card.

New SSD Drive For OS - How Should I Config My Drive Scheme Now?

I built a new PC in July based on ASUS P8P67 Pro Motherboard & i7-2600k 3.4 GHz processor w/ Cooler Master V8 heatsink. 16 GB RAM DDR3 Kingston HyperX. ASUS GTX 570 GeForce. Windows 7.
In stock (not OC) configuration, this computer should cruise pretty well. Instead, it is unstable and performance unimpressive. Premiere Pro CS 5.5 is sluggish with frequent timeline lockups and slow response to media export commands to AME and from with Premiere. After Effects too slow to be practical. Encore frequently crashes. I reboot several times in a typical editing day. I had similar problems with my older computer - that's why I built a new one.
I tried to follow all the tips in this forum for hard drive setup, etc. Benchmarking shows decent times for exports. The problem is, the computer freezes for a while before it responds to the export command, so any speed is negated. The entire Premiere screen grays out, says "not responding" for a minute, then finally AME opens up. Another minute goes by before it starts doing its thing.
I reinstalled CS and Windows several times. Uninstalled Matrox MX02 Mini and its software. Same problem.
I suspected my C Drive could be bad because I frequently get failure messages. Utilities like Norton System Works and System Mechanic Pro almost daily said my registry had errors and drives needed realignment or defrag (OS drive usually)
Having said all that, I went ahead and got an SSD. I now have the following:
120 GB SSD Kingston HyperX - NEW replacement for 1 TB Barracuda SATA 7200.12
2 TB SATA 3 Seagate Barracuda XT - Previews & Pagefile
2 TB SATA 3 Seagate Barracuda XT - Capture Video/Audio
1.5 TB SATA 7200.11 - empty
1.5 TB SATA 7200.11 - Temp OS Drive (Windows 7 and Programs)
2 TB G RAID External - Completed Projects
2 TB NAS Drive - Exported Videos & Backups
I have room for 5 internal hard drives and ample 850 Watts power
I plan on a clean install of Windows rather than cloning the OS drive. What other considerations are important here? Hard drive settings? Why is my C drive partitioned with a reserve?
According to previous tips, pagefile should go on fastest drive, previously one of the SATA 3s. Should it now now go on the SSD?
Should I RAID I have not attempted to overclock until I can get it running smoothly.

Percoplus,
Suggest:
1) Do a Windows format on your SSD before you load Windows; that way you can skip the strange partition that Win7 install wants to put on a "fresh" drive
2) You will get much better CS5.5 performance with 2 RAID 0 arrays than with individual drives:
- configure 2 2 drive arrays with you matching drives
- remove Windows "indexing" from both arrays; set drive hardware policy to "enable write caching" and "turn off Windows write-cache buffer flushing" on both arrays; test both arrays for sustained reads and writes using a utility such as HD Tune Pro (set block size to 2MB) and make sure you are getting at least 100MB/sec performance from both arrays for reads and writes; if you are not, figure out whether you have a driver or drive issue and resolve
3) format the arrays using GPT partitioning and 4k cluster size
4) put projects/media on the fastest array; put scratch, media cache, media cache DB, and export to the other array
Your system should NOT be sluggish, unless you have something wrong at this point!
Test with PPBM5 and make sure results line up with similar cpu/drive systems.
Regarding where to put Windows pagefile, it doesn't really matter too much since you have 16GB of RAM and Windows will not be needing the pagefile for much. I'd probably put it on the slower RAID array.
Regards,
Jim

Aperture Advice: Managed or Referenced Files

I've looked through a lot of the threads here and done some googling. It seems the answer is - "it depends".
So I was hoping I could get some advice on how I can maxamise my use of aperture.
System specs - Mac OSX Imac 2.66 GHz dual core. 4gb ddr2 Ram.
Part of the problem I'm having is that I have a huge number of photos to manage (it's around 350gb of photos) so because of this i'm keeping my photos stored on a fast(ish) 1 terrabyte external drive. At the moment i'm using a managed system but I'm told part of the pro's of aperture is that it's easy to change your file systems.
I don't have any need to edit my image masters, if i need to access photos i just export a version.
So am i right in thinking the best system for me, will be to keep a library file on my Imac but set managed storage to keep the masters on my terrabyte external drive?
Would really appreciate some advice on how I can best use aperture i've noticed that whenever it's open it's a total system hog. I basically have to close it if I want to use anything else on the Imac. Final cut express and aperture open is a no-no for example.
Any advice is greatly appreciated
Thanks
C
(Detailed system specs)
Model Name:          iMac
Model Identifier:          iMac8,1
Processor Name:          Intel Core 2 Duo
Processor Speed:          2.66 GHz
Number Of Processors:          1
Total Number Of Cores:          2
L2 Cache:          6 MB
Memory:          4 GB
Bus Speed:          1.07 GHz
Boot ROM Version:          IM81.00C1.B00
SMC Version (system):          1.29f1

Why managed, I am lazy! One click back up using vaults, simplicity.
I have multiple RAIDs, my old "slow" RAID now has my libraries on it, the new "fast" RAIDs are used for video.
I do run multiple libraries and the only time I have had an issue with missing files was when playing with referenced. That said many people enjoy running a referenced system, for the way I work managed is easiest. Disk space is not an issue, backing up is not an issue, what is it now, $100 for a 2TB drive, on site and offsite covered for $200.
Tony

3 or 4 questions... or more!!! :)

Hi!
does the fan speed control works or not?????
Where can I find the modded bios for the raid promise controler? Is it better???? (benchmarks).
Does Msi and Promise have new driver and/or patch / bios that solves "slow raid performance" 85/55 hdbench... (slow compared against other competitors that uses the same controler...)
Anyone have sometimes on boot up, when hd are spinning, a clicking noise???( head parking) like the one made when the computer is turned off???
Is the MSI gf4 ti4200 is a great product???? noisy video hs fan???? Is she realy run hot after 20 min of gaming???? I want to buy one... any advise is great!!! The other choise was the Leadtek ... the only thing I find interesting about leadtek is its huge Heatsink...
Thanks!!

Hello
You can find the Raid Performance Patch at viaarena.com.
As far as the noise your HD is making, mine did that about a week before it went out on me. It lasted just long enough to get my data off of it. It started when I was shutting down, then after awhile it did it on boot, then it just started doing it all the time.

RAID0+1 on the MSI7220 (K8N Diamond Plus)

Well, with a little good fortune I was able to acquire 3 extra hard drives, and after a week and a half of fiddling with the connections and benchmarking the hell out of them with IOMeter, I have a blazing fast RAID0+1 setup for speed and reliability!
(take breath)
I'll serve up some dirt (the details) after I make it look pretty, but here are the quick skinny numbers. First off, using IOMeter requires an advanced degree in rocket science, so I basicallly stole some numbers from tweakers.net and ran the access specification for all tests with a "file server" profile using 100% random access, 80% read/20% write, with 60% overall being 4K IO, 10% each of 512B and 64K, and the rest evenly divided at 4% of 1K, 2K, 8K, 16K and 32K. I tried various iterations of using a single drive (tested each of the four nVidia ports, and each of the two Silicon Image ports), then the drives in pairs, then for the nVidia, in triples and finally all four. Then I moved to using the RAID drivers and tested each of the above combinations in configurations of RAID0 (striping), RAID1 (mirroring), RAID5 (distributed striping with distributed parity) and finally RAID0+1 (mirror of stripes).
And the winners are (dah dah dah dahhhhh):
eXtreme raw IO performance with NO redundancy: RAID0 with 4 disks (which isn't surprising as the more spindles you get, the faster the RAID0).
A *very* close second in performance but WITH redundancy (albeit half the capacity): RAID0+1 with 4 disks.
A sort of close runner up with redundancy and a dead compromise in capacity: RAID5 with 4 disks.
Just in case you were wondering.

In general you are correct.
With 4 total disks, and RAID 0, you have the raw capacity of all 4 drives, but your risk of losing the volume is 4 times greater than a single drive, because any 1 of the disks failing will cause you to lose the whole array. It does yield the fastest performance though for both reading and writing, because if, for example, you use a 32K stripe, and write a 128K file to the disk, 32K parts of the file are written to each disk simultaneously, and same for reading.
With all 4 disks in RAID 5, you effectively lose 1 whole disk to parity, and the result is the raw capacity of 3 drives, but your risk for failure is cut in half, because you would need 2 drives to fail concurrently to lose the volume (1 drive can go down and the system will operate in a degraded state-- slightly slower because it has to calculate the missing data from parity, and effectively becomes a very slow RAID 0 until the failed disk is restored). The write performance is very slow (relatively speaking) because there is the extra calculation of the parity bit, but read performance is very fast.
With all 4 in a RAID 0+1, you lose 2 disks since logically you are making two 2-disk stripes, and then mirroring them. You end up with the raw capacity of 2 drives, but you have slightly better fault tolerance than the RAID 5, depending on your luck. If you lose 1 disk, that stripe is effectively down, so the mirrored stripe is "active", and you can still lose the second disk in the failed stripe without losing data. However, if a disk in the "active" stripe fails concurrently, you would lose the array. The performance drop in the degraded state yields a 2-disk stripe performance, which is still double the performance of a single disk. The "normal" performance of a RAID0+1 is very good, with write performance being equivalent to a single disk, but read performance being that of a 4-disk stripe (because the reads are distributed across the mirrors, and then across the stripes, 2x2 improvement). It is only marginally slower than RAID0, but with the benefit of redundancy.
The best stripe size is arbitrary and depends on the makeup of your overall file system. I'll post more on that later today, right now I have to catch a train.

Raid mode very slow/ crashing after installing driver (Z68A-GD65 (G3))

I recently purchased a Z68A-GD65 (G3) motherboard, i5-2500K, 8 gb ram, and a Crucial 64GB 4m SSD drive. The intention was to use the SSD drive for Intel Smart Response Tech.
Instructions tell me to place the SATA ports in RAID mode, which I did. During installation of Windows 7 (64-bit), I provided the "F6" raid drivers which installed. After that point, my hard drive (a WD 250GB SATAII drive) begins to perform incredibly slow. Once Windows 7 installs, which takes way longer than it should, Windows has problems booting and won't work properly.
I tried installing Windows 7 without the "F6" raid drive on install and that worked fine. Everything seemed to be working great. I then installed the Intel Rapid Storage Tech software and rebooted. Now my system is acting up again. It takes far longer to boot, it shuts down as soon as it boots, and it has even blue screened on boot. It is very unstable.
My SSD is unplugged during all of this. I'm just trying to get Windows 7 to run properly on my one mechanical drive in raid mode.
Do I have a bad motherboard? Anyone have any ideas?
Edit: I'm not using the Marvell ports. The hard drive works fine in IDE/ACHI modes... I've been using it for years.
Thanks

So it turned out that the issue was the hard drive itself. It had been running flawlessly for years as my system drive. However, the Intel iastor driver exposed a hidden issue with the drive.
I decided to check the Windows system event log for any clues. In there I saw errors with event id 9 on iastor. They were lined up with the times of the hangs and blue screens. Solutions I found online for this issue did not help. Luckily I have two hard drives in my system, although they are both the same model. I moved my data around and clean installed Windows on the other drive instead. The hangs and blue screens are no longer occurring.

Windows Server 2012 Storage Spaces Simple RAID 0 VERY SLOW reads, but fast writes with LSI 9207-8e SAS JBOD HBA Controller

Has anyone else seen Windows Server 2012 Storage Spaces with a Simple RAID 0 (also happens with Mirrored RAID 1 and Parity RAID 5) virtual disk exhibiting extremely slow read speed of 5Mb/sec, yet write performance is normal at 650Mb/sec in RAID 0?
Windows Server 2012 Standard
Intel i7 CPU and Motherboard
LSI 9207-8e 6Gb SAS JBOD Controller with latest firmware/BIOS and Windows driver.
(4) Hitachi 4TB 6Gb SATA Enterprise Hard Disk Drives HUS724040ALE640
(4) Hitachi 4TB 6Gb SATA Desktop Hard Disk Drives HDS724040ALE640
Hitachi drives are directly connected to LSI 9207-8e using a 2-meter SAS SFF-8088 to eSATA cable to six-inch eSATA/SATA adapter.
The Enterprise drives are on LSI's compatibility list. The Desktop drives are not, but regardless, both drive models are affected by the problem.
Interestingly, this entire configuration but with two SIIG eSATA 2-Port adapters instead of the LSI 9207-8e, works perfectly with both reads and writes at 670Mb/sec.
I thought SAS was going to be a sure bet for expanding beyond the capacity of port limited eSATA adapters, but after a week of frustration and spending over $5,000.00 on drives, controllers and cabling, it's time to ask for help!
Any similar experiences or solutions?

Has anyone else seen Windows Server 2012 Storage Spaces with a Simple RAID 0 (also happens with Mirrored RAID 1 and Parity RAID 5) virtual disk exhibiting extremely slow read speed of 5Mb/sec, yet write performance is normal at 650Mb/sec in RAID 0?
Windows Server 2012 Standard
Intel i7 CPU and Motherboard
LSI 9207-8e 6Gb SAS JBOD Controller with latest firmware/BIOS and Windows driver.
(4) Hitachi 4TB 6Gb SATA Enterprise Hard Disk Drives HUS724040ALE640
(4) Hitachi 4TB 6Gb SATA Desktop Hard Disk Drives HDS724040ALE640
Hitachi drives are directly connected to LSI 9207-8e using a 2-meter SAS SFF-8088 to eSATA cable to six-inch eSATA/SATA adapter.
The Enterprise drives are on LSI's compatibility list. The Desktop drives are not, but regardless, both drive models are affected by the problem.
Interestingly, this entire configuration but with two SIIG eSATA 2-Port adapters instead of the LSI 9207-8e, works perfectly with both reads and writes at 670Mb/sec.
I thought SAS was going to be a sure bet for expanding beyond the capacity of port limited eSATA adapters, but after a week of frustration and spending over $5,000.00 on drives, controllers and cabling, it's time to ask for help!
Any similar experiences or solutions?
1) Yes, being slow either on reads or on writes is a quite common situation for storage spaces. See references (with some of the solutions I hope):
http://social.technet.microsoft.com/Forums/en-US/winserverfiles/thread/a58f8fce-de45-4032-a3ef-f825ee39b96e/
http://blogs.technet.com/b/askpfeplat/archive/2012/10/10/windows-server-2012-storage-spaces-is-it-for-you-could-be.aspx
http://social.technet.microsoft.com/Forums/en-US/winserver8gen/thread/64aff15f-2e34-40c6-a873-2e0da5a355d2/
and this one is my favorite putting a lot of light on the issue:
http://helgeklein.com/blog/2012/03/windows-8-storage-spaces-bugs-and-design-flaws/
2) Issues with SATA-to-SAS hardware is also very common. See:
http://social.technet.microsoft.com/Forums/en-US/winserverClustering/thread/5d4f68b7-5fc4-4a3c-8232-a2a68bf3e6d2
StarWind iSCSI SAN & NAS

Disk Utility restore on a RAID is extremely slow

here is my story: i want to replace old internal 250gb hdd with 2х500gb striped raid set.
i have 2 G5 Powermacs, so i put 2х500 in one of them and build striped raid. then i booted it in target disk mode and connected via firewire to other G5. other G5 was booted from system dvd. i launched disk utility and choose restore from old 250 hd to new raid set. there are 100 gb of data on 250 hdd.
for now this process goes about 5 hours, and blue line is just on a half. is it normal when i restore a single partition on a raid, or something is wrong? why it's sooo slow?

No, changing/adding them to a Raid will lose all info.
Haven't actually tried it on a Concantated RAID, but I still suspect the Directory structure to be so different that it will wipe any info.
Only way is to backup everthing on the HD then move it back after it's setup for RAID.
RAID is really nice for the speed,but boy is it ever unreliable, of all the RAID setups I've had, the longest one went before going bonkers is 6 months, most far less. Never rovered one of them either.

HP ENVY 15: Slow 4K read speed for Dual SSD's in RAID 0 - Needs to be addressed!

Crystalmark scores 2nd gen Envy 15:
Read/Write
SEQ 352 / 215
512K 270 / 195
4K 14.2 / 110
http://forum.notebookreview.com/showthread.php?t=427278&page=754
read write
418.2 / 213.9 seq
251.3 / 181.1 random 512k
9.8 / 53.17 random 4k
http://forum.notebookreview.com/showthread.php?t=427278&page=750
4k Read Speed MUCH SLOWER than without Raid 0 (~20-25 mb/s)

Not exactly sure, I pulled those benchmarks off of notebookreview. All I know from reading the forum is that HP Envy 15's in RAID 0 seem to be severely underperforming 4k read speeds relative to the same SSD's not in RAID 0. The 4k random read is the MOST IMPORTANT benchmark for real world performance.
Other posts in the old HP forum have cited that disabled C1E could possibly fix the 4k read issue, however the Envy 15 seems to have no way to access this option in the BIOS.

Slow Raid

Similar Messages

Maybe you are looking for