I have been looking to swap my current 4 Western Digital WD10EACS00ZJB0 1TB hard drives out for some 2 TB, but I am concerned about all the DOA, "died within 3 months", "drive clicks", "locks up", and "%40 sectors are bad" reviews I keep seeing. At least %30-%50 are bad. For all drives across the board.
I have had my current drives online 24/7 for close to 2 years, and I fear they are going to fail at anytime.
It seems like my keeping them awake, and spinning 24/7, must be part of the reason they are still alive right now, so I plan on treating the new drives the same way, but I see how new WD pull lots of load cycles due to some 8 second hardware based spin down. Also I would like to stay away from anything that "adjusts" my spin speed.
Are there any drives, specific model numbers, that can hold up for my current XP-hacked, software based RAID 5 setup?
I use them for storing video, and serve up separate files over Wifi to at least 4 devices at once. I like my video to serve up quick, so I never let them spin down.
I was thinking about the new Drobo FS, but I need some hard drives first.
Just looking for reliable drives, that will last as long as my current ones have.
In addition to TLER, the "RE" drives have an unrecoverable read error rate of only 1 in 10^15 bits read which is going to be important for a RAID-5 array. I've seen posts of people trying to build multi-terabyte RAID arrays out of drives with 1 in 10^14 bit error rates, and at those rates there's a better than even chance that a 10TB array will be unrecoverable after a drive failure.
Most of WD's standard "Green" drives that I've checked also have 1 in 10^15 bit error rates, although they don't have TLER.
April 6, 2010 9:56:17 PM
Is the only difference between the RE and non RE the TLER? If so would using the WDTLR.exe make them equal, if I'm not all that concerned with any performance increase?
Do you know if the "RAID read issue" applies to software RAIDs as well?
Would using a Drobo FS in RAID 6 mode eliminate the need to worry about the read error issue, since the likely-hood of the read error being in the same place astronomical?
Sorry for all the questions, but $600+ dollars for the drives alone is a lot of money, not to mention the possible loss of all my data.
The RE drives are classified as "Enterprise" products so they're built to a higher standard than the non-RE drives. One reflection of that is the better unrecoverable read error rate (for the non-green versions), but it also shows up in other specs such as twice as many load/unload cycles.
In terms of RAID functionality with the WDTLER utility I'd expect them to be essentially the same. But WD no longer supports the utility and they say that it may not be compatible with the firmware on newer drives. That may just be FUD, but unless you can find someone who's had success using it with a specific model you've purchased there's always a chance you could end up bricking your drive.
If by "RAID read issue" you mean the problem of being able to recover from the failure of a large array when the chance of an unrecoverable read error is so high then yes - it affects all RAID-5 arrays whether they're hardware, software, or some combination thereof.
RAID-6 is much safer in that respect as the only way you can fail to recover is if you have unrecoverable read errors at exactly the same sectors on multiple drives - a very unlikely prospect. But a lot of software RAID doesn't support RAID-6, and for software RAID it can be more CPU-intensive than RAID-5.