|
Volguus posted:I've never had a problem with seagate for the last ... more than a decade or so. Still have their hdd's running more than 10 years later in various computers around the house. I did, however, have my company buying 100 WD hdd's (15 years ago, but still), and 80 of them died within the month. All of them were gone in 6 months. As I understand it, high failure rates like this are typically the result of a localized event - more than likely the box the drives were shipped in suffered a severe impact (say, getting dropped by a deliveryman), resulting in loose dust, misaligned platters, etc. Not enough damage for the drives to fail outright, but enough that after a period of time the drives would begin to exhibit failure symptoms and be summarily pulled from service. This is also why even businesses that order drives in bulk aren't necessarily qualified to offer evidence with regards to failure rates - unless you're constantly picking up new drives on different days from different delivery services, you're still at the mercy of localized events skewing your data.
|
#
?
Oct 14, 2014 10:48
|
|
|
|
| # ? Jun 7, 2024 22:50 |
|
|
EVO drives are affected. Pro drives are not. Which one is it? There's no "EVO Pro" drive.
|
|
#
?
Oct 14, 2014 12:33
|
|
|
Factory Factory posted:EVO drives are affected. Pro drives are not. Which one is it? There's no "EVO Pro" drive. One of these I think. I can see that I made a mistake in the naming convention, but dude didn't buy one with packaging, he just got a bare drive 
|
|
#
?
Oct 14, 2014 12:38
|
|
|
How sure are we that the firmware is going to be released tomorrow? Any updates, test reports?
|
|
#
?
Oct 14, 2014 15:30
|
|
|
r0ck0 posted:How sure are we that the firmware is going to be released tomorrow? Any updates, test reports? That's when Samsung said it would be coming.
|
|
#
?
Oct 14, 2014 15:40
|
|
|
I wouldn't be surprised if it wasn't exactly the 15th that it ends up coming out on either.
|
|
#
?
Oct 14, 2014 16:06
|
|
|
My old Intel drive got hit by the 8MB bug. I noticed while reading about it that the firmware update that was supposed to fix it didn't actually fix it. Did they ever actually fix it or what?
|
|
#
?
Oct 14, 2014 16:11
|
|
|
wdarkk posted:My old Intel drive got hit by the 8MB bug. I noticed while reading about it that the firmware update that was supposed to fix it didn't actually fix it. Did they ever actually fix it or what? They fixed the circumstances for the original bug, unexpected power offs, but it seems that the 8 MB thing is the failure case for the drive on every problem. Bright side: they won't hesitate to RMA it.
|
|
#
?
Oct 14, 2014 17:44
|
|
|
isndl posted:This is also why even businesses that order drives in bulk aren't necessarily qualified to offer evidence with regards to failure rates - unless you're constantly picking up new drives on different days from different delivery services, you're still at the mercy of localized events skewing your data. Indeed, better to have nothing at all and just go by anecdote. I think some data on failure rates is better than no data especially when these aren't scientific publications, but maybe we just agree to disagree there.
|
|
#
?
Oct 14, 2014 23:34
|
|
|
Rakthar posted:Indeed, better to have nothing at all and just go by anecdote. I don't know if what he's saying is true, but if it is true, then what he's describing is an anecdote.
|
|
#
?
Oct 15, 2014 01:08
|
|
|
Thermopyle posted:I don't know if what he's saying is true, but if it is true, then what he's describing is an anecdote. I find the reports that backblaze and google have put out to contain very useful information. Some people look at those reports and bring up concerns over the data and mention aspects like bulk purchasing, different vendors, and drive farming as reasons it is not reliable. I disagree and feel that the data they do present is better than not having it.
|
|
#
?
Oct 15, 2014 01:17
|
|
|
Rakthar posted:I find the reports that backblaze and google have put out to contain very useful information. Some people look at those reports and bring up concerns over the data and mention aspects like bulk purchasing, different vendors, and drive farming as reasons it is not reliable. I disagree and feel that the data they do present is better than not having it. I don't think he was talking about backblaze and google. He said "unless you're constantly picking up new drives on different days from different delivery services". I'm sure google and backblaze get lots of drives from different lots, delivery services, etc.
|
|
#
?
Oct 15, 2014 01:22
|
|
|
Didn't one of them actually include delivery mechanism as a metric? I think it was backblaze since they had to get creative with acquisitions during the monsoon aftermath.
|
|
#
?
Oct 15, 2014 01:27
|
|
|
Tech Report has done some preliminary testing of the updated 840 EVO firmware, and it seems to work as expected and is non-destructive, though it may take some time (it mentions a 500GB model taking over an hour, but doesn't mention how much data was on the SSD).
|
|
#
?
Oct 15, 2014 01:31
|
|
|
GokieKS posted:Tech Report has done some preliminary testing of the updated 840 EVO firmware, and it seems to work as expected and is non-destructive, though it may take some time (it mentions a 500GB model taking over an hour, but doesn't mention how much data was on the SSD). The most interesting quote in that article: "For what it's worth, Samsung says the problem is unique to the 840 EVO and not an artifact of that drive's TLC NAND. The older, TLC-based 840 Series is unaffected, and so is the upcoming 850 EVO, which will use the three-bit version of 3D V-NAND."
|
|
#
?
Oct 15, 2014 01:59
|
|
|
So the "fix" is some workaround poo poo as expected. Rewrite the entire drive once so the data is fresh, then the firmware will continually rewrite old data in the background so that it remains fresh. E: There's a more indepth article here: http://www.pcper.com/reviews/Storage/Samsung-840-EVO-Performance-Restoration-Tool-preview-Getting-EVOs-back-speed/Update- Sphyre fucked around with this message at 02:31 on Oct 15, 2014 |
|
#
?
Oct 15, 2014 02:19
|
|
|
Sphyre posted:So the "fix" is some workaround poo poo as expected. Rewrite the entire drive once so the data is fresh, then the firmware will continually rewrite old data in the background so that it remains fresh.
|
|
#
?
Oct 15, 2014 02:33
|
|
|
Alereon posted:To expand a bit, this seems to confirm that the issue was with the DSP algorithm Samsung was using to read data. Anandtech has a bit of information on using DSP techniques to read data from NAND in their Samsung 840 review. It wasn't calibrating itself correctly, so wasn't reacting to changing cell levels as it should have been. They are fixing this calibration and re-writing fresh, recalibrated data. This statement from Samsung: Samsung posted:This only occurs if the data was kept in its initial cell without changing, and there are no symptoms of reduced read performance if the data was subsequently migrated from those cells or overwritten. In other words, as the SSD is used more and more over time, the performance decrease disappears naturally. For those who want to solve the issue quickly, this software restores the read performance by rewriting the old data. The time taken to complete the procedure depends on the amount of data stored. Doesn't mention anything about a corrected read algorithm or rewriting the old data to new calibrated values, but does mention that the performance would disappear "naturally" (as the new firmware rewrites data on the drive periodically) and that the initial rewrite is done to expedite this process. I'm guessing that the performance of these drives would continue to degrade if left unpowered for a couple of weeks.
|
|
#
?
Oct 15, 2014 02:56
|
|
|
Sphyre posted:Doesn't mention anything about a corrected read algorithm or rewriting the old data to new calibrated values, but does mention that the performance would disappear "naturally" (as the new firmware rewrites data on the drive periodically) and that the initial rewrite is done to expedite this process. I'm guessing that the performance of these drives would continue to degrade if left unpowered for a couple of weeks. Samsung posted:SSDs usually calibrate changes in the statuses of cells over time via the flash management software algorithm. Due to the error in the software algorithm, the 840 EVO performed read-retry processes aggressively, resulting in a drop in overall read performance. This only occurs if the data was kept in its initial cell without changing, and there are no symptoms of reduced read performance if the data was subsequently migrated from those cells or overwritten.
|
|
#
?
Oct 15, 2014 03:06
|
|
|
Volguus posted:HGST is just Hitachi's hdd section bought by WD. So, most likely (for now) they have different everything. Question is ... for how long?
|
|
#
?
Oct 15, 2014 03:35
|
|
|
deimos posted:Didn't one of them actually include delivery mechanism as a metric? I think it was backblaze since they had to get creative with acquisitions during the monsoon aftermath. Yeah, they were buying external 3.5" disks and hiring temp workers to shuck them. Teams driving vans around to the big-box stores buying as many external disks as they had. The kinda nutty stuff that makes me wish I worked for them.  Alereon posted:The corrected read algorithm is the firmware update we are applying, the Performance Restoration Wizard is re-writing the recalibrated data. I think Sphyre's contention is that the firmware is going to itself rewrite the data on a periodic basis (in the background), causing a slow erosion of write cycles even on a disk with no write activity. It's not really clear if the fix is on the DSP side (better able to handle the aging of data) or on the wear-leveling side (more aggressivly rewriting older cells to refresh them)
|
|
#
?
Oct 15, 2014 07:47
|
|
|
Harik posted:I think Sphyre's contention is that the firmware is going to itself rewrite the data on a periodic basis (in the background), causing a slow erosion of write cycles even on a disk with no write activity. It's not really clear if the fix is on the DSP side (better able to handle the aging of data) or on the wear-leveling side (more aggressivly rewriting older cells to refresh them) Yep, that's exactly it. The performance restoration software has been released: https://www.samsung.com/global/business/semiconductor/minisite/SSD/us/html/support/downloads.html
|
|
#
?
Oct 15, 2014 10:33
|
|
|
Ran the Samsung performance restoration software and compared to HDtach result ran last month: 
|
|
#
?
Oct 15, 2014 12:53
|
|
|
When they said the 'performance decrease would disappear naturally' it probably means the data will be refreshed naturally through wear leveling, while the tool they provide you can do it in one pass. And in the future this wouldn't happen or needed in the first place because the charge value are correctly calibrated with updated firmware and wont cause the DSP to retry aggressively and kill the reading speed during old reads. DaNzA fucked around with this message at 13:00 on Oct 15, 2014 |
|
#
?
Oct 15, 2014 12:58
|
|
|
quote:10)PC will be shut down automatically after firmware update(Performance Restoration counts down 20 seconds before shutdown) Does this mean if I do this update remotely and do not have a way to turn the PC back on it will stay off?
|
|
#
?
Oct 15, 2014 15:17
|
|
|
r0ck0 posted:Does this mean if I do this update remotely and do not have a way to turn the PC back on it will stay off? Yes. It shuts down, not restarts (doing it now)
|
|
#
?
Oct 15, 2014 15:21
|
|
|
r0ck0 posted:Does this mean if I do this update remotely and do not have a way to turn the PC back on it will stay off? It did for mine. Though in hindsight I'm not entirely sure I have Teamviewer setup to start on boot. Edit: Oh, well, there you go.
|
|
#
?
Oct 15, 2014 15:22
|
|
|
WhiskeyJuvenile posted:Yes. It shuts down, not restarts (doing it now) Well ain't that some bullshit. I just did it and now I have to wait till I get home tonight to continue. edit. Now is the first time I wish I had WOL setup. I should have setup WOL first, its too early for this poo poo. r0ck0 fucked around with this message at 15:36 on Oct 15, 2014 |
|
#
?
Oct 15, 2014 15:22
|
|
|
What's the verdict on the fix/update? Is it risky wrt. potential data loss? Would the option for making a full drive backup, preferably to some sort of image file on a spare drive, before attempting to apply the patch, again be Macrium Reflect?
|
|
#
?
Oct 15, 2014 18:08
|
|
|
What is the advantage of using Macrium Reflect vs. System Restore in Windows? My typical process to clone to a new SSD has always been: 1) Backup with System Restore 2) Switch out SSD's, install Windows on new one (my case doesn't have room for both at once) 3) Restore backup to new SSD 4) Expand partition size for larger SSD Does Macrium save me some steps in this process?
|
|
#
?
Oct 15, 2014 18:41
|
|
|
nielsm posted:What's the verdict on the fix/update? Is it risky wrt. potential data loss? I just ran the update on a 250 and 500. It worked fine, no problems besides shutting the computer down after the firmware update. edit: shutting the computer down is a normal step, it just annoyed me.
|
|
#
?
Oct 15, 2014 18:42
|
|
|
BitterAvatar posted:What is the advantage of using Macrium Reflect vs. System Restore in Windows? It clones and resizes for you.
|
|
#
?
Oct 15, 2014 18:53
|
|
|
BitterAvatar posted:What is the advantage of using Macrium Reflect vs. System Restore in Windows? It's one step instead of four, and handles all the partition sizing automagically.
|
|
#
?
Oct 15, 2014 19:27
|
|
|
...and doesn't involve installing Windows.
|
|
#
?
Oct 15, 2014 19:30
|
|
|
Schpyder posted:It's one step instead of four, and handles all the partition sizing automagically. Well that does sound very useful compared to Windows. Cat Hatter posted:...and doesn't involve installing Windows. It will let me backup the OS, pull out the existing SSD, and replace it with the new one? How does it handle this?
|
|
#
?
Oct 15, 2014 19:58
|
|
|
BitterAvatar posted:Well that does sound very useful compared to Windows. You have to have both old and new drives installed simultaneously.
|
|
#
?
Oct 15, 2014 20:07
|
|
|
Why can't you have the old drive dangling outside the case? What does the case being too small have to do with it?
|
|
#
?
Oct 15, 2014 20:16
|
|
|
Buy a $5 USB to SATA cable or a $15 USB 3.0 to SATA enclosure
|
|
#
?
Oct 15, 2014 20:22
|
|
|
Updated 2 500GB EVOs and everything went flawlessly. Seemed like it would take around 40 mins to do the after-the-reboot performance leveling. Good to go!
|
|
#
?
Oct 15, 2014 20:39
|
|
|
|
| # ? Jun 7, 2024 22:50 |
|
|
Before: After:  Only took an hour to do my 1tb EVO, but I only got like 170gb on it.
|
|
#
?
Oct 15, 2014 20:49
|
|
