Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
WD Red NAS drives and ZFS
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 19035

PostPosted: Wed Jul 21, 2021 5:33 am    Post subject: Reply with quote

szatox wrote:
Quote:
I agree that Seagate is probably the brand I've seen fail most, but also that was due to it being the most used.
Yeah, the cheapest drives in the market, right?
I have no idea if they are the cheapest. When I've made purchases in the last couple of years, they have not been. But my failure comment was about enterprise drives where the server vendor handles sourcing. Three would occasionally not be a Seagate, but I don't recall the brand(s).

szatox wrote:
I've heard people complaining about their failures, but I've been quite happy with their performance.
I think when people have a bad experience it can bias their future decisions. I stopped buying Seagate when they purchased Conner because they mixed those drives into existing product lines without differentiation. I didn't have an actual failure, but I stopped trusting them at that point.
Conner wrote:
In the mid-1990s, just prior to their buyout by Seagate, Conner Peripherals started using a cost-saving measure on its low-end drives. Instead of bolted-down, cast aluminum lids with rubber gaskets, Conner Peripherals came up with a design that used a thinner stamped aluminum lid, and a thick adhesive tape seal along the perimeter of the lid, where the lip of the lid meets the base casting. This design used no screws to hold on the lid — just cutting the tape could permanently damage the drive by making the lid come off. This design was kept well after the Seagate buyout, and was discontinued in 1999.
Had they made a clear distinction, I'd probably have bought at least some of their drives. I probably went from Maxtor to WD to Samsung.


Moved from Documentation, Tips & Tricks to Gentoo Chat.
_________________
Magna Carta (1215) | Spectral evidence no longer permissible (c. 1792) | Cancel culture, deplatforming (c. 2016)
Back to top
View user's profile Send private message
Goverp
Veteran
Veteran


Joined: 07 Mar 2007
Posts: 1199

PostPosted: Wed Jul 21, 2021 8:19 am    Post subject: Reply with quote

FWIW my current desktop uses Toshiba 1TB drives. It's 18 mths old, set up with an 5-disk RAID 10 array. No problems so far, and fast.
While researching, I came across the drawbacks of shingled drives, which is why I chose 1TB.
I also studied https://www.userbenchmark.com, and Tosh drives seemed about the best at the time.
_________________
Greybeard
Back to top
View user's profile Send private message
eccerr0r
Watchman
Watchman


Joined: 01 Jul 2004
Posts: 8164
Location: almost Mile High in the USA

PostPosted: Wed Jul 21, 2021 1:28 pm    Post subject: Reply with quote

I have two 2TB disks on death watch now, one with 1000 reallocated blocks and another that's spewing random read errors, former is a Toshiba, latter is a... WD. MDRAID has been able to deal with/rewrite the bad blocks enough so that they don't get kicked...yet...

TBH I have a WD "Green" as a single drive --
Code:
  5 Reallocated_Sector_Ct   0x0033   200   200   140    Pre-fail  Always       -       0
  9 Power_On_Hours          0x0032   001   001   000    Old_age   Always       -       72683
193 Load_Cycle_Count        0x0032   001   001   000    Old_age   Always       -       628752
196 Reallocated_Event_Count 0x0032   200   200   000    Old_age   Always       -       0
197 Current_Pending_Sector  0x0032   200   200   000    Old_age   Always       -       0

Not sure if this should also be on death watch, but it's been reliable, no read/write failures yet. However the "WORST" fields for the advisory fields are "very close" to the failing mark...
_________________
Intel Core i7 2700K/RadeonHD 5770/24GB DDR3/256GB SSD
What am I supposed watching?
Back to top
View user's profile Send private message
Jaglover
Watchman
Watchman


Joined: 29 May 2005
Posts: 8043
Location: Saint Amant, Acadiana

PostPosted: Wed Jul 21, 2021 2:03 pm    Post subject: Reply with quote

I have one Green, about same age as yours.
Code:
  5 Reallocated_Sector_Ct   0x0033   200   200   140    Pre-fail  Always       -       0
  9 Power_On_Hours          0x0032   005   005   000    Old_age   Always       -       69441
 12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       -       75
193 Load_Cycle_Count        0x0032   200   200   000    Old_age   Always       -       30

See the diff in Load Cycle Count?
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!
Back to top
View user's profile Send private message
eccerr0r
Watchman
Watchman


Joined: 01 Jul 2004
Posts: 8164
Location: almost Mile High in the USA

PostPosted: Wed Jul 21, 2021 3:39 pm    Post subject: Reply with quote

Yeah, I stopped that load cycling fairly late. However it's still working... It's in a PVR box that has suffered through other hardware failures (motherboard, PSU)... with all the things together I'm surprised it's still working.
_________________
Intel Core i7 2700K/RadeonHD 5770/24GB DDR3/256GB SSD
What am I supposed watching?
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 19035

PostPosted: Wed Jul 21, 2021 6:34 pm    Post subject: Reply with quote

Goverp wrote:
FWIW my current desktop uses Toshiba 1TB drives. It's 18 mths old, set up with an 5-disk RAID 10 array. No problems so far, and fast.
While researching, I came across the drawbacks of shingled drives, which is why I chose 1TB.
I also studied https://www.userbenchmark.com, and Tosh drives seemed about the best at the time.
I may go back to 1TB. One main reason for 2TB was convenience and being able to move partitions around, but 1TB is probably enough. If I'm not misremembering, the prices were pretty close between 1 and 2 TB. That would leave the 4 TB drives for a file server with at least one spare, and the 2TB drives could be for temporary data that isn't a concern in the event of a loss.


eccerr0r wrote:
I have two 2TB disks on death watch now, one with 1000 reallocated blocks and another that's spewing random read errors, former is a Toshiba, latter is a... WD. MDRAID has been able to deal with/rewrite the bad blocks enough so that they don't get kicked...yet...
My concern isn't with drives that are old, rather drives that fail earlier than a reasonable life expectancy. The Samsung drive I'm replacing only had ~74k power on hours, but it started to cause some compile failures. For whatever difference it makes, the load cycle count was 120 around that time.

The first two (gcc, llvm on different dates) compiled after retrying. I thought maybe it was heat, so I adjusted MAKEOPTS. Another llvm upgrade failed and then again with a retry. I think the failure was something about an "internal compiler error". Anyway, searching led to indications of hardware issues, which is when I noticed the drive issues.
_________________
Magna Carta (1215) | Spectral evidence no longer permissible (c. 1792) | Cancel culture, deplatforming (c. 2016)
Back to top
View user's profile Send private message
eccerr0r
Watchman
Watchman


Joined: 01 Jul 2004
Posts: 8164
Location: almost Mile High in the USA

PostPosted: Wed Jul 21, 2021 8:19 pm    Post subject: Reply with quote

Well, knock on wood - the bad toshiba is one of three in the array. The other two look okay, just that one.

Wish buying any random 2T drive would get a CMR, alas this will not be the case *sigh*
_________________
Intel Core i7 2700K/RadeonHD 5770/24GB DDR3/256GB SSD
What am I supposed watching?
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 5437
Location: Dallas area

PostPosted: Wed Jul 21, 2021 8:28 pm    Post subject: Reply with quote

Was curious so just looked at amazon wd red nas vs plus and for 2/3 TB they're the same price (more or less)
Plus runs more for 4TB and up though.
And the smr's only seem to go to 6 tb, while the plus line runs to 14tb.

Quote:
According to WD:
The company is taking a new branding initiative to clarify the technology used in its NAS drives—in the near future, "WD Red" will exclusively mean disks using Shingled Magnetic Recording technology, and "WD Red Plus" will mean disks using Conventional Magnetic Recording.


But even with reasonable prices, I don't trust WD, they've already played fast and loose with facts, and that puts me off.

This from the middle of last year so not sure how current https://nascompares.com/answer/list-of-wd-cmr-and-smr-hard-drives-hdd/
_________________
PRIME x570-pro, 3700x, 5.11 zen kernel
gcc 9.3.0/11.1.1, profile 17.1 amd64-no-multilib, openrc, wayland

copy of my local repo
---
The New OTW
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 19035

PostPosted: Wed Jul 21, 2021 11:35 pm    Post subject: Reply with quote

With sales I've only paid ~$10 more since the SMR change. If I read correctly, I was only seeing about $5 difference between WD 4 TB SMR and CMR for the non sale prices.
_________________
Magna Carta (1215) | Spectral evidence no longer permissible (c. 1792) | Cancel culture, deplatforming (c. 2016)
Back to top
View user's profile Send private message
Jaglover
Watchman
Watchman


Joined: 29 May 2005
Posts: 8043
Location: Saint Amant, Acadiana

PostPosted: Sat Jul 24, 2021 1:26 am    Post subject: Reply with quote

I started this thread when one drive was dropped from zpool. It is replaced now and I've been playing with this drive. It didn't pass smartctl test and had pending bad sectors. I ran badblocks in write mode on it. Now it has no bad sectors, no pending bad sectors, no reallocated sectors, passes smartctl test. Can someone smarter tell me what happened to it?

Code:
Device: sat16:/dev/sdc
Type: 16 Byte SCSI ATA SAT Passthru
Size: 2861588 MiB
Model: [WDC WD30EFRX-68AX9N0]
Serial: [WD-WMC1T3032314]
Firmware: [80.00A80]
SMART Available: yes
Quirks:
Awake: yes
SMART Disk Health Good: yes
Off-line Data Collection Status: [Off-line data collection activity was never started.]
Total Time To Complete Off-Line Data Collection: 38880 s
Self-Test Execution Status: [Self-test routine in progress]
Percent Self-Test Remaining: 30%
Conveyance Self-Test Available: yes
Short/Extended Self-Test Available: yes
Start Self-Test Available: yes
Abort Self-Test Available: yes
Short Self-Test Polling Time: 2 min
Extended Self-Test Polling Time: 390 min
Conveyance Self-Test Polling Time: 5 min
Bad Sectors: 0 sectors
Powered On: 2.5 years
Power Cycles: 26
Average Powered On Per Power Cycle: 1.2 months
Temperature: 44.0 C
Attribute Parsing Verification: Good
Overall Status: GOOD
ID# Name                        Value Worst Thres Pretty      Raw            Type    Updates Good Good/Past
  1 raw-read-error-rate         200   200    51   36          0x240000000000 prefail online  yes  yes
  3 spin-up-time                193   176    21   5.3 s       0xcd1400000000 prefail online  yes  yes
  4 start-stop-count            100   100     0   26          0x1a0000000000 old-age online  n/a  n/a
  5 reallocated-sector-count    200   200   140   0 sectors   0x000000000000 prefail online  yes  yes
  7 seek-error-rate             200   199     0   0           0x000000000000 old-age online  n/a  n/a
  9 power-on-hours               70    70     0   2.5 years   0x0d5600000000 old-age online  n/a  n/a
 10 spin-retry-count            100   253     0   0           0x000000000000 old-age online  n/a  n/a
 11 calibration-retry-count     100   253     0   0           0x000000000000 old-age online  n/a  n/a
 12 power-cycle-count           100   100     0   26          0x1a0000000000 old-age online  n/a  n/a
192 power-off-retract-count     200   200     0   11          0x0b0000000000 old-age online  n/a  n/a
193 load-cycle-count            200   200     0   14          0x0e0000000000 old-age online  n/a  n/a
194 temperature-celsius-2       106   101     0   40.0 C      0x2c0000000000 old-age online  n/a  n/a
196 reallocated-event-count     200   200     0   0           0x000000000000 old-age online  n/a  n/a
197 current-pending-sector      200   200     0   0 sectors   0x000000000000 old-age online  n/a  n/a
198 offline-uncorrectable       100   253     0   0 sectors   0x000000000000 old-age offline n/a  n/a
199 udma-crc-error-count        200   200     0   0           0x000000000000 old-age online  n/a  n/a
200 multi-zone-error-rate       200   200     0   12          0x0c0000000000 old-age offline n/a  n/a

_________________
My Gentoo installation notes.
Please learn how to denote units correctly!
Back to top
View user's profile Send private message
Goverp
Veteran
Veteran


Joined: 07 Mar 2007
Posts: 1199

PostPosted: Sat Jul 24, 2021 10:13 am    Post subject: Reply with quote

Code:
Self-Test Execution Status: [Self-test routine in progress]
Percent Self-Test Remaining: 30%

_________________
Greybeard
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 5437
Location: Dallas area

PostPosted: Sat Jul 24, 2021 10:15 am    Post subject: Reply with quote

Jaglover wrote:
I started this thread when one drive was dropped from zpool. It is replaced now and I've been playing with this drive. It didn't pass smartctl test and had pending bad sectors. I ran badblocks in write mode on it. Now it has no bad sectors, no pending bad sectors, no reallocated sectors, passes smartctl test. Can someone smarter tell me what happened to it?


I've had spinning rust come back to life after a low level format, ie write to whole disk using vendor software, running badblocks w/write has a similar effect.
Though I never fully trust the drive after that, I use it, but only for very temp storage, and if it fails ... oh well. :lol:
I have had them in use for quite a while after a low level format for some drives.

I think what's going on is the disk "firmware" gets out of sync with the real hardware, and when a low level format is done, it resets it all to a known state, ie both firmware and the disk itself agree.

The only real drawback to low level formattting is it takes time. :)
_________________
PRIME x570-pro, 3700x, 5.11 zen kernel
gcc 9.3.0/11.1.1, profile 17.1 amd64-no-multilib, openrc, wayland

copy of my local repo
---
The New OTW
Back to top
View user's profile Send private message
steve_v
Apprentice
Apprentice


Joined: 20 Jun 2004
Posts: 259
Location: New Zealand

PostPosted: Sat Jul 24, 2021 10:18 am    Post subject: Reply with quote

IME modern disk firmware can also "change it's mind" about failing sectors if it gets a few clean read-write cycles there. It's possible that's what happened here, but I'd definitely wait until it has passed an extended SMART selftest (or two) before coming to any conclusions. I've also seen drives where the "reallocated sector count" is pure fiction, but that was a long time ago.

Personally I quite like hdrecover for the purposes of smacking the firmware upside the head when it sits there diddling about with "pending sectors", and it's a bit faster than writing over the whole drive too.

If you want to get properly low-level and see what's up there's always MHDD... But that's a whole other kettle of fish. Also most of the documentation is in Russian :D
_________________
Once is happenstance. Twice is coincidence. Three times is enemy action. Four times is Official GNOME Policy.
Back to top
View user's profile Send private message
Jaglover
Watchman
Watchman


Joined: 29 May 2005
Posts: 8043
Location: Saint Amant, Acadiana

PostPosted: Sat Jul 24, 2021 10:55 am    Post subject: Reply with quote

Second long test is completed, no bad sectors.
MHDD is new to me. I remember when there was low level format in BIOS menu and SCSI software, didn't know one still can do it at home.
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!
Back to top
View user's profile Send private message
steve_v
Apprentice
Apprentice


Joined: 20 Jun 2004
Posts: 259
Location: New Zealand

PostPosted: Sat Jul 24, 2021 11:17 am    Post subject: Reply with quote

Jaglover wrote:
MHDD is new to me.

MHDD is very old, very low-level, and very, very dangerous.
If you've never tried to diagnose/recover a drive from real-mode DOS or felt the need to have a face-to-chipset chat with a drive controller, I'm not surprised you've never heard of it.

It's only now returned to the top of my mental stack because I was using it yesterday to gently prod an old 850MB WD Caviar with a stuck head actuator back into life. Complete success there BTW, with some fascinating (from a pure curiosity point of view) neurophysics data from 1995 as a prize. :)

Even if you're not planning on sending low-level ATA commands manually, MHDDs scan and dump functions can be handy for identifying sectors that aren't truly "bad" but still exhibit anomalous behaviour - such as occasionally taking several times longer to read than they should. That's exactly the kind of thing that will have the firmware marking and unmarking them as they hover around it's "this sector no good" remap threshold.
Then again, it's old, and there are probably more modern equivalents with better support for modern drives.

If you want to really get below the firmware layer on a modern drive, you'll almost certainly need special cables and expensive (also difficult to acquire) software. Such software does exist, but I haven't got it.
_________________
Once is happenstance. Twice is coincidence. Three times is enemy action. Four times is Official GNOME Policy.
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 48619
Location: 56N 3W

PostPosted: Sat Jul 24, 2021 2:08 pm    Post subject: Reply with quote

Jaglover,

Correct. It's not possible to low level format voice coil head movement drives at home.
You may recall that users trashed drives in the early days voice coil head movement drives by attempting the low level format.
The firmware was quickly fixed ta that the drives went through the motions but internally did nothing.
That kept all the software commanding low level format happy too.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
eccerr0r
Watchman
Watchman


Joined: 01 Jul 2004
Posts: 8164
Location: almost Mile High in the USA

PostPosted: Sat Jul 24, 2021 3:25 pm    Post subject: Reply with quote

What I'm curious of is if there's a way to "forget" all the grown/remapped bad blocks and retest them from scratch.

Seems like even this requires special software.
_________________
Intel Core i7 2700K/RadeonHD 5770/24GB DDR3/256GB SSD
What am I supposed watching?
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 5437
Location: Dallas area

PostPosted: Sat Jul 24, 2021 4:00 pm    Post subject: Reply with quote

eccerr0r wrote:
What I'm curious of is if there's a way to "forget" all the grown/remapped bad blocks and retest them from scratch.

Seems like even this requires special software.


I used to download the low level stuff from disk manf, when I would buy one of their products, early stuff all ran under dos no matter which disk manf.
It pretty much starts from scratch and writes zeros in every sector, then tries to read them and then marks them accordingly.
I' ve done this with seagate, wd, maxtor and hgst.

The bad blocks program will do the same. It tries a sector and sets/unsets the bb indicator appropriately.

Edit to add: When I say low level, I mean low level formatting, not low level hardware interface, ie talking to disk controller.
_________________
PRIME x570-pro, 3700x, 5.11 zen kernel
gcc 9.3.0/11.1.1, profile 17.1 amd64-no-multilib, openrc, wayland

copy of my local repo
---
The New OTW
Back to top
View user's profile Send private message
eccerr0r
Watchman
Watchman


Joined: 01 Jul 2004
Posts: 8164
Location: almost Mile High in the USA

PostPosted: Sat Jul 24, 2021 5:14 pm    Post subject: Reply with quote

There's a distinction here that's not quite clear:
when I mean reset, I mean reset. Currently when you run badblocks(8) or even a lot of manufacturer diagnostic tools, it will write to the remapped block, not to the original block that was remapped - so this is not exactly what I mean.
_________________
Intel Core i7 2700K/RadeonHD 5770/24GB DDR3/256GB SSD
What am I supposed watching?
Back to top
View user's profile Send private message
Jaglover
Watchman
Watchman


Joined: 29 May 2005
Posts: 8043
Location: Saint Amant, Acadiana

PostPosted: Sat Jul 24, 2021 5:25 pm    Post subject: Reply with quote

Somewhere in my memory cells is a piece of info I heard once, no proof it is correct. Anyhow, someone alleged first few bad sectors will be remapped without being counted. I had 4 bad sectors waiting remapping at some point, they vanished without a trace. My idea to run badblocks in write mode was only meant to force the drive to do something about them. Well, it certainly did ... but what?
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!
Back to top
View user's profile Send private message
steve_v
Apprentice
Apprentice


Joined: 20 Jun 2004
Posts: 259
Location: New Zealand

PostPosted: Sat Jul 24, 2021 6:00 pm    Post subject: Reply with quote

eccerr0r wrote:
Currently when you run badblocks(8) or even a lot of manufacturer diagnostic tools, it will write to the remapped block, not to the original block that was remapped

Indeed. I'm pretty sure you can't get at the reserved areas or make the firmware "unremap" sectors without insider software though, and likely some mucking about with jtag ports to boot.
That's secret manufacturer territory, and we're really not supposed to go there.

Jaglover wrote:
Somewhere in my memory cells is a piece of info I heard once, no proof it is correct. Anyhow, someone alleged first few bad sectors will be remapped without being counted.

I have heard that too, but It was a long time ago and have no reliable source. It may well have just been some bit of nonsense from a fanboy bashing $other_brand for masking defects, so I didn't mention it.
_________________
Once is happenstance. Twice is coincidence. Three times is enemy action. Four times is Official GNOME Policy.
Back to top
View user's profile Send private message
Goverp
Veteran
Veteran


Joined: 07 Mar 2007
Posts: 1199

PostPosted: Sat Jul 24, 2021 6:51 pm    Post subject: Reply with quote

I read that secure erase comes in two flavours, one that really, really, overwrites everything. Presumably that would at least ignore the sector remapping, so maybe it might restore them? The second link mentions "restore DCO" which might be the thing, but seems only worth trying if you would otherwise junk the drive.
https://ata.wiki.kernel.org/index.php/ATA_Secure_Erase
https://tinyapps.org/docs/wipe_drives_hdparm.html#n4
_________________
Greybeard
Back to top
View user's profile Send private message
szatox
Advocate
Advocate


Joined: 27 Aug 2013
Posts: 2089

PostPosted: Sun Jul 25, 2021 9:38 am    Post subject: Reply with quote

I can add 2 flavors more:
- Encrypting the drive before using it - enables quick wipe by destroying the key.
- Hitting it with a hammer/magnet/microwaves/sanding the plate. (Warning: you won't be able to use this drive again)
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2
Page 2 of 2

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum