The GetDPI Photography Forum
Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!
MacPro RAID Card
- Thread starter Bob
- Start date
Lars
Active member
Thanks, good to know what to expect from CS3.
I'm prepping the remaining disks in my RAID for re-use. I did recover some of the data in between read errors, but it was a rather painful process that I'd rather not repeat - nothing can make me trust that card again. Got myself an extra drive for onsite backup now (incidentally the same WD 500GB with 16MB cache, it was €120 here on Malta which is not too shabby considering the remote location).
Being a bit worried by the recent experience with my RAID controller I decided to use Windows built-in support for striped and mirrored partitions. The advantage of using Windows stripe/mirror is that the disk format will be independent of the disk controller used, so if my system fails I can hook up the array to any controller or any Windows-based desktop system. There might be a slight performance penalty compared to an ideal hardware RAID controller.
However being a bit offsite here and with a tight budget there are limits to how much I can safeguard and verify. An onsite backup will of course make sleep better.
I'm prepping the remaining disks in my RAID for re-use. I did recover some of the data in between read errors, but it was a rather painful process that I'd rather not repeat - nothing can make me trust that card again. Got myself an extra drive for onsite backup now (incidentally the same WD 500GB with 16MB cache, it was €120 here on Malta which is not too shabby considering the remote location).
Being a bit worried by the recent experience with my RAID controller I decided to use Windows built-in support for striped and mirrored partitions. The advantage of using Windows stripe/mirror is that the disk format will be independent of the disk controller used, so if my system fails I can hook up the array to any controller or any Windows-based desktop system. There might be a slight performance penalty compared to an ideal hardware RAID controller.
However being a bit offsite here and with a tight budget there are limits to how much I can safeguard and verify. An onsite backup will of course make sleep better.
Lars,
Hmmm WD Drives.
Originally, I loaded my MacPro with four Hitachi 500GB HDT7250VLA360 drives which are available to me at a pretty good price. I also am using a lot of these drives in other applications, including in RAIDS from other suppliers and getting really good results. I have not spent/wasted the time to fully debug this, but three of the four new drives failed in the raid within a few days. and at this point cannot be read at all by the Apple hardware. I replaced all of my $100.00 drives with Apple's $250.00 versions. Apple uses Seagate ST3500630AS P drives. I do not know if the trailing "P" is significant and cannot find it listed in Seagate's online catalog.
The s=Seagate drives work well and continue to work well. What seems to have happened to the Hitachi drives is that they have had their interface re-configured possibly be a spurious configuration command. This is a one way trip, and cannot be un-done unless you happen to haver a interface board that supports the proper variant, which the Apple RAID seems to not. I complained to Apple that there might be a firmware or OS bug causing this problem, but they were not impressed and claimed that they only support those drives that Apple sells. The Seagate drive does not have the same low level configuration commands as the Hitachi drives.
So, if you see multiple drive failures all at once and you are using Hitachi drives, then you might be seeing a repeat of my experience.
-bob
Hmmm WD Drives.
Originally, I loaded my MacPro with four Hitachi 500GB HDT7250VLA360 drives which are available to me at a pretty good price. I also am using a lot of these drives in other applications, including in RAIDS from other suppliers and getting really good results. I have not spent/wasted the time to fully debug this, but three of the four new drives failed in the raid within a few days. and at this point cannot be read at all by the Apple hardware. I replaced all of my $100.00 drives with Apple's $250.00 versions. Apple uses Seagate ST3500630AS P drives. I do not know if the trailing "P" is significant and cannot find it listed in Seagate's online catalog.
The s=Seagate drives work well and continue to work well. What seems to have happened to the Hitachi drives is that they have had their interface re-configured possibly be a spurious configuration command. This is a one way trip, and cannot be un-done unless you happen to haver a interface board that supports the proper variant, which the Apple RAID seems to not. I complained to Apple that there might be a firmware or OS bug causing this problem, but they were not impressed and claimed that they only support those drives that Apple sells. The Seagate drive does not have the same low level configuration commands as the Hitachi drives.
So, if you see multiple drive failures all at once and you are using Hitachi drives, then you might be seeing a repeat of my experience.
-bob
Lars
Active member
Hehe my drives are Seagate 
Barracuda 7200.9 ST33000622AS.
After reading about Google's experience with drives I doubt one manufacturer is better than others. There seems to be more variation in reliability between manufacturing runs - sometimes you're lucky sometimes not.
Re Apple's expensive drives, possibly with a higher price point they can afford to determine the quality of a specific series. It's also possible that Apple's drives are RAID edition drives - retries and timeouts reconfigured to not disable the array in case of a read error. Those drives cost a bit more. The the rest of the price difference is directly correlated to Apple's stock price OEM parts are always at a premium, in all industries from cars to computers to watches.
Still the lesson here is that hard drives are to be considered as reliable as floppies - sometimes they fail. Google's stats were interesting, over 20% of drives fail within 3 years. Manufacturers' MTBF numbers are grossly exaggerated. http://research.google.com/archive/disk_failures.pdf
Barracuda 7200.9 ST33000622AS.After reading about Google's experience with drives I doubt one manufacturer is better than others. There seems to be more variation in reliability between manufacturing runs - sometimes you're lucky sometimes not.
Re Apple's expensive drives, possibly with a higher price point they can afford to determine the quality of a specific series. It's also possible that Apple's drives are RAID edition drives - retries and timeouts reconfigured to not disable the array in case of a read error. Those drives cost a bit more. The the rest of the price difference is directly correlated to Apple's stock price
OEM parts are always at a premium, in all industries from cars to computers to watches.Still the lesson here is that hard drives are to be considered as reliable as floppies - sometimes they fail. Google's stats were interesting, over 20% of drives fail within 3 years. Manufacturers' MTBF numbers are grossly exaggerated. http://research.google.com/archive/disk_failures.pdf
On scratch: If you set efficiency to show on your info screen and monitor it, any time CS3 gets below about 95% you can assume it's starting to scratch. If it drops to 65%, it's scratching a lot for that image --- it's the most reliable way I've found to determine I'm actually using the scratch disk for scratching...
On drives: Historically, I preferred Hitachi or IBM drives. Then a few years ago, I found Seagates to be very reliable. I had bad experiences with a few WD drives early on -- mostly heat and noise related, not failure related -- so avoided them. Recently, WD released the Green series 1 TB drive (32 MB buffer) so I bought one for a remote quad-interface box (eSATA, FW800, FW400, USB2) to port all of my images on PC disks over to Mac format disks. Bottom line was the WD was impressive! VERY fast, quiet and never got more than warm in the NON-cooled enclosure during the sustained write and subsequent reads of 500 MB blocks of data from and to the other drives. These are still a bit pricey per Gig at $279 from www.macsales.com.
Re Seagate: Seagate just introduced a 7200.11 drive that has shown impressive increases in sustained read and write speeds over the previous version drive, the 7200.10. Both the older generation .10 and the new .11 drives use "perpendicular" technology which improves efficiency, so either of these drives is a pretty hefty step up from the 7200.9, but the .11 is second generation perpendicular with 32 MB buffer (up form 16 on the .10). Right now, the 7200.11's are pricey at $149 (macsales) for a 500, or about on par with the 1 TB drive cost/gig. However, my local Fry's just put palette loads of 7200.10 500's on the floors at $99, IMO a great deal for a lower-cost RAID array or basic data back-up...
On drives: Historically, I preferred Hitachi or IBM drives. Then a few years ago, I found Seagates to be very reliable. I had bad experiences with a few WD drives early on -- mostly heat and noise related, not failure related -- so avoided them. Recently, WD released the Green series 1 TB drive (32 MB buffer) so I bought one for a remote quad-interface box (eSATA, FW800, FW400, USB2) to port all of my images on PC disks over to Mac format disks. Bottom line was the WD was impressive! VERY fast, quiet and never got more than warm in the NON-cooled enclosure during the sustained write and subsequent reads of 500 MB blocks of data from and to the other drives. These are still a bit pricey per Gig at $279 from www.macsales.com.
Re Seagate: Seagate just introduced a 7200.11 drive that has shown impressive increases in sustained read and write speeds over the previous version drive, the 7200.10. Both the older generation .10 and the new .11 drives use "perpendicular" technology which improves efficiency, so either of these drives is a pretty hefty step up from the 7200.9, but the .11 is second generation perpendicular with 32 MB buffer (up form 16 on the .10). Right now, the 7200.11's are pricey at $149 (macsales) for a 500, or about on par with the 1 TB drive cost/gig. However, my local Fry's just put palette loads of 7200.10 500's on the floors at $99, IMO a great deal for a lower-cost RAID array or basic data back-up...
PS: On my new Mac Pro, I am (correction will be as soon as it arrives --- hopefully tomorrow!) using an Apple OEM Seagate 500G for the OS (don't know which it is yet, but with the $100 upgrade cost over the 320, I'm hoping it's a .11). I have a second Seagate 500G 7200.11 that will be partitioned 150/350, the 150 "fast" sector reserved for scratch, the remaining 350 critical back-up with OS drive image. Next slot will be a second WD 1TB for images (mirrored to the current one), last slot a Seagate 500G 7200.10 for Time Machine.
Since drives are now relatively inexpensive, my back-up routine has been redundant RAID 1-1; a fully redundant copy onsite, then a second redundant copy offsite. I use older 500G drives for offsite back-up of current and historical images. these are just bare drives which I connect with a simple USB2 to IDE/SATA adapter: http://eshop.macsales.com/item/Newer Technology/U2NVSPATARH/ --- I just store the bare drives in a bubble pack in a safe and update them monthly or after any large project. Call me paranoid, but I am a big fan of multiple redundancy for critical data like my images.
Since drives are now relatively inexpensive, my back-up routine has been redundant RAID 1-1; a fully redundant copy onsite, then a second redundant copy offsite. I use older 500G drives for offsite back-up of current and historical images. these are just bare drives which I connect with a simple USB2 to IDE/SATA adapter: http://eshop.macsales.com/item/Newer Technology/U2NVSPATARH/ --- I just store the bare drives in a bubble pack in a safe and update them monthly or after any large project. Call me paranoid, but I am a big fan of multiple redundancy for critical data like my images.
Daniel
New member
my two past nightmare experiences made me "paranoid" too. now that most drives are relatively inexpensive, it'd be daft not to have at least one extra drive for emergency recovery.
as far as data safety goes, i subscribe to the idea of wearing suspenders and belts to hold up my fitted pants. i have two redundant mirror configuration in my mac pro, two external redundant mirror configuration, another external redundant drive, another drive stored in a lock-box along with the same files backed-up to DVDs. one external drive is designated as my quick-grab drive if there's an emergency - such as fire.
i like my suspenders and belts very much. thank you
PS: there was an excellent article in the NYTIMES dated December 23, 2007 article, The Afterlife Is Expensive for Digital Movies, by Michael Ciepy about the expense of archiving digital media. it's worth reading.
http://www.nytimes.com/2007/12/23/b...e+Is+Expensive+for+Digital+Movies&oref=slogin