Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Anyone here using an OWC Thunderbay enclosure?
- Thread starter odinsride
- Start date
- Sort by reaction score
FYI, there was an update to SoftRAID 5 today, - version 5.0.6. Apparently fixes a critical RAID 5 rebuild bug, and it's recommended everyone update.
Boy do I NOT feel good when those sorts of updates come out...
I suppose, I feel better being on 5.0.6 now.
WilliamG ... are you by chance using SoftRAID 5 on the same machine that you have Windows BootCamp installed? If so, have you noticed any "issues" with Windows crashing about 2 minutes after logging in (assuming your Thunderbay is connected and powered on)? I realize the SoftRAID is not being used by Windows, but its mere presence seems to be at the root of the issue for me. SoftRAID 0 is no problem ... only SoftRAID 5.
At first I was convinced it was the Thunderbay IV causing the problems, but now I suspect SoftRAID 5 has something in the boot sectors on the disks that Windows eventually gets around to checking and doesn't like.
Last edited:
I don't like the new higher price, but if you are buying a OWC Thunderbay enclosure at the same time you can get it cheaper.
That being said ... I do like it because of the disk monitoring and email notifications of pending or actual problems, especially if it is on a remote system like a Mini Server which is normally unattended. SoftRAID also comes with some test and diagnostic programs that I find useful, especially with a brand new untested disk drive.
So yes, I would prefer SoftRAID in RAID 0 or 1 rather than using the simple Disk Utility Apple software RAID.
Unfortunately, I can't offer any useful info here. It's a 2012 Mac mini running the OWCs, and my 5K iMac running Boot Camp. Have you tried uninstalling SoftRAID to see if that solves the issue?
I second hfg. Apple's built in RAID software is basic, and even using that word does a disservice to basic software everywhere. SoftRAID is the real deal. That said, if all you're looking for is basic RAID 0 with no monitoring etc, it'll do the job.
I don't want to completely derail this thread .... but a quick summary:
RE: Thunderbay IV and SoftRAID 5 causes Windows BootCamp to crash
I created a test platform with Thunderbay IV and 3 new Seagate drives, plus a Thunderbolt Windows SSD. I used SoftRAID under OS X to format the 3 drives as zeroed-out-blank, RAID-1, RAID-0, and RAID-5 and rebooted to Windows (again zero-out the drives between each test). All but the RAID-5 configuration had no issues when Windows ran. Only the RAID-5 caused Windows to crash, oddly enough, about 2 minutes after login. It didn't matter if Windows was external on Thunderbolt, or standard internal BootCamp install.
I have since also tested with a Pegasus J4 enclosure of 4 2.5" drives on Thunderbolt. This requires a OS X driver to even be visible to OS X, but it still seems to screw up Windows when the disks are RAID-5 formatted.
I have tried to solve this with OWC and SoftRAID tech support without success. I suspect there is some data pattern on the boot sectors of the RAID-5 formatted disks that Windows doesn't like, but it takes some time for Windows to examine them and as a result ... crash.
Turn off the Thunderbay, and Windows runs fine ... but not a solution if you daisy-chain the Thunderbolt, or wish to have the Windows SSD inside the Thunderbay (as I now have, but the disks are not RAID-5).
-howard
That doesn't sound like fun at all. Hopefully it's something you'll be able to work around, as that sort of thing is unlikely (I'd guess) be fixed anytime soon, if ever. It's one of those things that will likely affect only the few, and I know what it's like to be one of those few, historically!
This entire thread has been very helpful! I've been traveling like crazy since receiving my OWC Thunderbay and I'm just now getting ready to format and partition my RAID 5 setup. (I've already certified all of the drives, to include my cold spare.) In a quest to optimize performance, whether real-world or benchmarked numbers, I need to get some opinions from my fellow experts.
Here's my situation:
I'm a heavy-duty photographer that's also getting into a fair amount of video work.
Here's my set-up:
OWC Thunderbay 4 RAID 5 Edition w/ four 5 TB drives and one cold 5 TB spare drive.
Will back-up onsite to a Time Capsule and various other stand-alone hard drives.
I also backup to off-site drives once a month by bringing them home and then return them to their offsite location the next day.
Here are my questions/dilemma:
I want to to squeeze as much performance as I can out of my RAID 5. Digiloyd's excellent article here has given me the idea to partition my RAID 5 so that I utilize the faster sections of the drives for things like video and photo editing, and then use slower partitions on the drive for things that don't require speed, such as documents, music, etc. It's also beneficial as it keeps the partitioned volumes small enough so you can back them up to individual 5 TB drives.
So here's how I'm thinking about setting it up. What do you think? I know a lot of you have vast experience in this area and I would really appreciate your opinions.
Partition 1, 5 TB, fastest section of the drives: Video Footage Libraries (FCPX)
Partition 2, 5 TB, next fastest section of the drives: Photo Libraries (Currently Aperture, eventually migrating to Capture One Pro or Lightroom)
Partition 3, 4 TB, slowest section of the drives: Documents, iTunes Library, Finished movies, etc.
Thanks for the help/advice!
Bryan
Here's my situation:
I'm a heavy-duty photographer that's also getting into a fair amount of video work.
Here's my set-up:
OWC Thunderbay 4 RAID 5 Edition w/ four 5 TB drives and one cold 5 TB spare drive.
Will back-up onsite to a Time Capsule and various other stand-alone hard drives.
I also backup to off-site drives once a month by bringing them home and then return them to their offsite location the next day.
Here are my questions/dilemma:
I want to to squeeze as much performance as I can out of my RAID 5. Digiloyd's excellent article here has given me the idea to partition my RAID 5 so that I utilize the faster sections of the drives for things like video and photo editing, and then use slower partitions on the drive for things that don't require speed, such as documents, music, etc. It's also beneficial as it keeps the partitioned volumes small enough so you can back them up to individual 5 TB drives.
So here's how I'm thinking about setting it up. What do you think? I know a lot of you have vast experience in this area and I would really appreciate your opinions.
Partition 1, 5 TB, fastest section of the drives: Video Footage Libraries (FCPX)
Partition 2, 5 TB, next fastest section of the drives: Photo Libraries (Currently Aperture, eventually migrating to Capture One Pro or Lightroom)
Partition 3, 4 TB, slowest section of the drives: Documents, iTunes Library, Finished movies, etc.
Thanks for the help/advice!
Bryan
Last edited:
In regards to my post above, there are a lot of pros that have been posting in this thread and I'm hoping to draw out some opinions before I establish my RAID.
Do you think it would be smart to partition my RAID for even higher performance and potentially easier backup to single drives? (see my post above) I want to get this one right, even if it means using forward thinking to future-proof for speed.
Let's hear some advice!
Bryan
Do you think it would be smart to partition my RAID for even higher performance and potentially easier backup to single drives? (see my post above) I want to get this one right, even if it means using forward thinking to future-proof for speed.
Let's hear some advice!
Bryan
Thank you for the numbers. Very true, most people forget to account for the partitioning of the drives and of course, the loss of space due to the parity data in a RAID 5 setup.
I'm utilizing four 5 TB drives, so I anticipate the actual available space to be approximately 13.8 TBs.
Bryan
Bryan-
I started off with:
Partition 1 - 500 GB PS scratch
Partition 2 - 2 TB user folder
Partitions 3 - all else
I am now looking to remove all the data off the RAID so I can reformat as a single partition and utilize folders instead. Didn't see any clear advantage to the above and, in fact, that user folder ended up causing LOTS of read/write activity and a lot of noise as well as being pretty slow. I think the RAID array is fine for holding big data but don't expect performance equal to SSD despite all of the benchmarking tests.
On a separate note, does anyone know how to keep the Thunderbay from sleeping the drives? I'm finding significant lag when accessing my photos via Lightroom and hearing the drives spin up before the photos can be scrolled through.
*Thunderbay 4 with four 4 TB HGST HG0S03664 drives
I started off with:
Partition 1 - 500 GB PS scratch
Partition 2 - 2 TB user folder
Partitions 3 - all else
I am now looking to remove all the data off the RAID so I can reformat as a single partition and utilize folders instead. Didn't see any clear advantage to the above and, in fact, that user folder ended up causing LOTS of read/write activity and a lot of noise as well as being pretty slow. I think the RAID array is fine for holding big data but don't expect performance equal to SSD despite all of the benchmarking tests.
On a separate note, does anyone know how to keep the Thunderbay from sleeping the drives? I'm finding significant lag when accessing my photos via Lightroom and hearing the drives spin up before the photos can be scrolled through.
*Thunderbay 4 with four 4 TB HGST HG0S03664 drives
We have a problem with 2 different thunderbays crashing an iMac (i think the kernel panic). After ejecting the TB4 and powering it off, the iMac crashes and restarts as soon as the thunderbolt cable is unplugged from the TB4. This has happened with 2 different thunderbays and happens whether the two thunderbays are daisy chained (the furthest in the chain being unplugged causes it or if only one attached and unplugged this way same thing happens).
I actually just tried this on another imac and this one had the kernel panic when I powered off the drive. Both imacs are late 2013 27" running 10.10.2
Any ideas?
I actually just tried this on another imac and this one had the kernel panic when I powered off the drive. Both imacs are late 2013 27" running 10.10.2
Any ideas?
I did some very rudimentary testing with my TB4 Mini and posted my results on page 7. Long story short, play with the stripe size and "purpose" and find a good baseline to use as your test before settling on a final configuration.
The partitioning idea is a good one and will, theoretically, net some additional speed. Additionally, a RAID 4 configuration should be faster than RAID 5 and offer very similar protection. It would be well worth your time to test RAID 4 vs RAID 5.
Regardless of the configuration that you end up with, I think that you will be satisfied with the speed. FWIW, I have a TB4 w/4 5400RPM 3TB Seagate drives in addition to the Mini (w/SSD) and get very acceptable performance from that device as well. I don't remember the exact numbers/configuration off the top of my head but I do see performance in the 650/700 range in RAID 4. Not to shabby for some cheap spindles (this is a temp solution that I'm using as scratch disk during my migration from Aperture to Lightroom, I'll be replacing the 3TB drives with some higher performance disk once I'm finished with my migration).
No ideas but I can concur. I see the exact same behavior with my TB4 on my nMP, iMac and rMBP. FWIW, I do not have any issues with my TB4 Mini. I wonder whats different between the TB4 and TB4 Mini???
That article is best minimally deceptive if not flat out wrong if read in the broadest terms.
" The first partition will be the fastest one, thus guaranteeing high speed. "
That isn't true over a variety of workloads. That is only true if the disk's head(s) are restricted just to the those outer tracks. In the context of his "Master" , "Archive 1" , "Archive 2" example, the "Master" is highest speed as long as almost never use "Archive 1" and "Archive 2". Once start to actually use two or more of those partitions the speed is dependent upon immediately preceding write/read.
This is another somewhat dubious idea suggested by that post that may "happen to work" in some restricted subset but are not a good rule of thumb to follow. There is a difference between "cloning" data and backing it up. General back ups contain multiple versions of files. Exactly matching media sizes will generally work for exact clones, but back up media should be larger (either virtually (e.g., multiple tapes ) or just physically larger (e.g., 5TB disk to back up 4TB disk) ).
Fixed, non mutated media files will clone rather well. ( Finished movies, exported photos in TIFF/DNG/RAW, static music files ).
A back-up program that doesn't allow the user to point at a folder/direction and say "just back up just this" is somewhat lacking. Partitions aren't necessary with the use of some reasonable discipline. ( periodically check to see if folder is getting too big. If so move some of that to "offline" archiving disk(s). Also checking to see how full back up is once done; over time will see limit coming. )
These two may "happen to work" if the FCPX and Aperture/Lightroom applications are always used exclusively ( one at a time ). It skews things so FCPX will get marginal benefit when it is up.
The performance bump here on Partition 1 is going to be rather minimal. The larger you make your "short stroked" ( limited to outer track) partition the slower it gets even if restrict access just to that partition. The initial 10% of a disk has some increase in speed merits (e.g. 500GB of a 5TB ) drive, but once get out around 15-20% you are not really buying all that much. The initial 5-7% would be better, so a 750-1,000 GB "Partition 1" would have substantial advantage.
Dividing into thirds isn't going to buy much in terms of performance.
The 2nd and 3rd partitions are highly dubious in terms of much performance difference. I'm sure it could be measured, but not sure it is makes a huge difference in user feel. To get the real user feel difference in "Partition 1" these two have to stuff that are extremely infrequently used ( stuff usually don't need). If once-in-a-blue-moon files where is the huge upside in small incrementally faster access time? "Faster than plugging in the offline archive and spinning it up" is the 'speed' metric that an online archive has to beat.
The flaw in the testing diagram in the article is that he is testing each partition one at a time. That isn't how they will be accessed real world if try to use them as three active data read/write targets.
Document archive ( don't intend to read any time soon)? Maybe. Documents folder that refer to and actively mutated? No.
Likewise with iTunes. .... Music almost never going to listen to? Maybe. iPhone/iPad back up storage? Yes, it is easy enough to schedule around working with other apps. Stuff that will play while working? No.
----------
Same general problems here.
The User folder contains the user preference and application support files in the Library folder. These will be accessed by foreground and some background apps ( notifications ). The periodic moves out of the Partition 1 into Partition 2-3 will nuke any speed increasing by "short stroking" the drive head(s) inside of Partition 1.
Archival ( don't touch except very rarely ) files are only thing that offer a performance advantage to "Partition 1".
In terms of back-up to a single drive, that can be done with a back-up application pointed at a single folder with a single drive destination.
P.S. the whole "User" folder in an external enclosure isn't generally a good idea. Minimally an "administration" user keep on the OS/Apps Volume is better so that can at least boot and do some maintenance if necessary. Non-admin users collected on an external volume is fine.
Last edited:
HFG - are you running Windows via Bootcamp with the Thunderbay? I picked one of these up last summer and since have found that it cannot be used (or even connected to a Mac) when running Bootcamp. I run Windows from an external SSD in a Seagate Thunderbolt sled, but when the TBIV is connected it causes Windows to hang as soon as it starts to boot up (before the Windows logo). Talking to OWC confirmed they are not supporting Bootcamp with the Thunderbay enclosures (pretty sure that extends to the newer 4 model). I'm exceptionally disappointed that I have to unplug the entire enclosure to run Bootcamp.
EDIT: Just read through the earlier posts and see your comments about Bootcamp. Still interested to hear if you were able to boot into Windows at all with the Thunderbay IV model, as I cannot.
As for crashing, I've not experienced this ever happening to me.
Last edited:
I just chatted with OWC support about my kernel panics and they say "the issue is a bug in 10.10.2 and the Apple Thunderbolt driver" and that apple is aware.
I have very mixed and confusing results when using the ThunderBay IV and Windows 8 BootCamp (either on the internal OS X drive or external ThunderBolt). I too received the same message from OWC that the ThunderBay does not support BootCamp.
However ... I do have a ThunderBay IV connected to a Retina iMac and the Windows 8 SSD is installed in Slot A of the ThunderBay and it works fine. The other slots have standard hard disks not controlled by SoftRAID. This is the setup I am currently using and typing this on.
Just as a test, I moved the Windows SSD to a Seagate GoFlex ThunderBolt sled and hooked a 4 drive ThunderBay IV running SoftRAID 5 to the Retina iMac and booted into Windows. To my surprise ... it worked just fine! I also talked to SoftRAID technical support (excellent support from Mark BTW!!) and he too was able to duplicate the setup with no problems. This also worked fine with the Windows SSD in a LaCie Rugged Thunderbolt enclosure.
I then moved the ThunderBay RAID 5 box over to a 2012 Mac Mini with Windows 8 BootCamp freshly installed internally on a shared 1TB SSD. Windows would not boot. I removed all the disks to just have the empty shell ... and Windows still would not boot. Unplug or power off the ThunderBay and it boots fine.
So ... yeah, I am still confused what the magic is to make it work all the time.
-howard
EDIT: One very subtle change to the SoftRAID 5 disks ... I have "named" them from the SoftRAID console which fixed a bug with the graphical "piping" showing the SoftRAID configuration. I don't know if this had any effect on making the SoftRAID 5 disks not crash Windows or not.
Last edited:
eoren1,
Thank you for the feedback -- much appreciated! Your post, and especially deconstruct60's post, really turned on the light bulb in my head. (More on this in a moment.) I'm very new to the ins and outs of storage administration, but am quickly learning thanks to the help I'm receiving here.
deconstruct60,
I can't thank you enough for responding with such an in-depth reply. You've really helped me understand that if the drive head is bouncing back and forth a lot between partitions, then any speed advantages are quickly lost and in some cases, performance may be even worse. (Sounds like eoren1 was experiencing this to some degree.)
Now that the light bulb has come on in my head, I am leaning towards partitioning the RAID into two pieces. The first 70% of the drives for nearly everything. (Video-editing, photo-editing, music, and more). Then using the last 30% of the drives for Partition 2. This section would only be used for PDFs, scanned documents that I've filed away, and finished standalone movies/video files that are rarely watched, and when they are, they are watched alone.
Essentially, keeping it very simple and using a normal folder structure inside the fast, first 70% of the drives...and then using the slower section for files that are rarely used. (And would still see speeds of 300 MB/s in a RAID 5 array.) This prevents the really slow part of the hard drives from being utilized for things that like speed, but at the same time, by utilizing it for others things, it keeps it from just being wasted space.
And by eliminating my original proposal of having 3 partitions, I avoid the drive heads moving back and forth between the partitions which would cause a potentially serious performance hit.
Am I on the right track now? Let me know and I'm going to get down to the business of setting this puppy up and migrating data.
Thanks,
Bryan
Bryan-
I would only use a partition (aka volume) to put a hard limit on the size of something (i.e. Time Machine).
Seriously, you have no idea the headache I am going through right now trying to undo the three partition setup I started with.
Just go with one large volume and use folders to separate out.
I use Carbon Copy Cloner to then keep backups on separate FW800 externals set to copy certain folders over.
E
I would only use a partition (aka volume) to put a hard limit on the size of something (i.e. Time Machine).
Seriously, you have no idea the headache I am going through right now trying to undo the three partition setup I started with.
Just go with one large volume and use folders to separate out.
I use Carbon Copy Cloner to then keep backups on separate FW800 externals set to copy certain folders over.
E
I have very mixed and confusing results when using the ThunderBay IV and Windows 8 BootCamp (either on the internal OS X drive or external ThunderBolt). I too received the same message from OWC that the ThunderBay does not support BootCamp.
However ... I do have a ThunderBay IV connected to a Retina iMac and the Windows 8 SSD is installed in Slot A of the ThunderBay and it works fine. The other slots have standard hard disks not controlled by SoftRAID. This is the setup I am currently using and typing this on.
Just as a test, I moved the Windows SSD to a Seagate GoFlex ThunderBolt sled and hooked a 4 drive ThunderBay IV running SoftRAID 5 to the Retina iMac and booted into Windows. To my surprise ... it worked just fine! I also talked to SoftRAID technical support (excellent support from Mark BTW!!) and he too was able to duplicate the setup with no problems. This also worked fine with the Windows SSD in a LaCie Rugged Thunderbolt enclosure.
I then moved the ThunderBay RAID 5 box over to a 2012 Mac Mini with Windows 8 BootCamp freshly installed internally on a shared 1TB SSD. Windows would not boot. I removed all the disks to just have the empty shell ... and Windows still would not boot. Unplug or power off the ThunderBay and it boots fine.
So ... yeah, I am still confused what the magic is to make it work all the time.
-howard
EDIT: One very subtle change to the SoftRAID 5 disks ... I have "named" them from the SoftRAID console which fixed a bug with the graphical "piping" showing the SoftRAID configuration. I don't know if this had any effect on making the SoftRAID 5 disks not crash Windows or not.
Interesting. So it sounds like the Thunderbolt 2 ports on the Retina iMac might make a difference? I had come from a Lacie 2big Thunderbolt external and had no issues running Windows in Bootcamp from that enclosure, so it was surprising and disappointing to find out the TB IV can't handle this, where the Lacie and the Seagate sled have no issues. Though the part that really annoys me is that I can't even leave it connected and boot from the Seagate sled.
I can't say that I really understand why the TB2 ports would be different, as the conversations I've had with OWC on this came back to something about TB chips they used if I recall correctly. But good to know there might be some use for this thing if I just upgrade...
E,
I hear you. I'm sure what you're going through is an incredible pain. I can also see why your existing setup did not work well and why you would want to change.
In my case though, can you think of a compelling reason a 10 TB volume and a 4 TB volume would be a bad idea on a 20 TB RAID 5 array? (Actual storage space is 14 TB.) My end goal is to keep the slow part of the drive separated from the rest, but of course, not at the expense of screwing something up either.
Sorry to hash this to death, I'm just trying to fully understand the ramifications of what I want to do. (And I'm ready to get things going!)
Thanks,
Bryan
E,
I hear you. I'm sure what you're going through is an incredible pain. I can also see why your existing setup did not work well and why you would want to change.
In my case though, can you think of a compelling reason a 10 TB volume and a 4 TB volume would be a bad idea on a 20 TB RAID 5 array? (Actual storage space is 14 TB.) My end goal is to keep the slow part of the drive separated from the rest, but of course, not at the expense of screwing something up either.
Sorry to hash this to death, I'm just trying to fully understand the ramifications of what I want to do. (And I'm ready to get things going!)
Thanks,
Bryan
I just don't think there is a real-world difference in speed if you cut up into one or two partitions and you are just limiting yourself with the latter.
If you were using the RAID for Time Machine, then absolutely you would want something to limit the amount of space used but otherwise I just don't see any reason for it.
Like I said before, the real world speeds of this using RAID5 are good but they are not SSD (or even crappy-single-SSD) equivalents. There is also a spin-up time when I access photos in LR that I'm trying to see if I can get around. All that is just to show that you are not going to get that theoretical 600-1,000 MB/sec when moving real data around day to day. Any difference in location of the data on the 'fastest' part of the drives would maybe amount to 5% at most (pulled that number out of thin air but still...)
One Volume - 'Snappier'?
So I finally was able to redo my Thunderbay 4 as a single 12TB volume (four 4TB HGST drives) instead of the 3 volume setup I had initially.
I also turned off the Energy Saver setting to put hard disks to sleep when possible.
Overally, the RAID5 array feels much snappier. Very responsive in all activities. I ran the DigiLloyd tools integrity checker on my photos which I had started using when one of my drives went bad and it read data at 450MB/sec over 2 TB. Not bad at all...
So I finally was able to redo my Thunderbay 4 as a single 12TB volume (four 4TB HGST drives) instead of the 3 volume setup I had initially.
I also turned off the Energy Saver setting to put hard disks to sleep when possible.
Overally, the RAID5 array feels much snappier. Very responsive in all activities. I ran the DigiLloyd tools integrity checker on my photos which I had started using when one of my drives went bad and it read data at 450MB/sec over 2 TB. Not bad at all...
I just got my Thunderbay 4. It came with a Zalman fan. I have it loaded with a SSD, two 3 tb drives, and a 1 tb drive. It's currently sitting on my desk, just next to my monitor.
I have to say, it's pretty loud Definitely not something that blends into the background. Some of it is the typical hard drive noises, but most of it is the fan.
I think short of putting it in a drawer, it's going to be loud.
I may consider the fan replacement.
Otherwise, it's been working fine and seems to be constructed pretty well.
I have to say, it's pretty loud Definitely not something that blends into the background. Some of it is the typical hard drive noises, but most of it is the fan.
I think short of putting it in a drawer, it's going to be loud.
I may consider the fan replacement.
Otherwise, it's been working fine and seems to be constructed pretty well.