Hard Disk Monitoring

I have started tests on my new Disk before trusting it with data - all its space is Unallocated - no partitions
Short Self-Test passed
Random Seek Test passed
Surface Test passed
Extended Self-Test is 30% complete after one hour.

Am I wasting my time ?
Should I have initialized as MBR and allocated space to partitions and formatted before testing ?

I have not seen any "Conveyance" Test Option.
Is this because nothing suitable has been defined for my specific Disk which is :-
Hard Disk Model ID . . . . . . . . . . . . . . . : WDC WD20EZRX-00D8PB0
Firmware Revision . . . . . . . . . . . . . . . : 80.00A80
Hard Disk Serial Number . . . . . . . . . . . . : WD-WCC4M2389833
Total Size . . . . . . . . . . . . . . . . . . . : 1907726 MB

When I started my career as a design engineer,
our products were calibrated and packed with 10 cms. of crumple packaging material between the case each each of the 6 sides of the shipping container,
and one sample would be dropped twice from a height of 3 feet onto each of the 6 sides and each of the 6 corners to ensure it would survive shipping.
and then unpacked and inspected for damage and then tested to determine if it was still within calibration limits.
I have seen railway porters with cardboard boxes marked by Remington as "Typewriter, delicate, handle with care",
transferring the unit from a trolley by drop kicking it into the goods wagon.

I was horrified when I opened the box from Amazon to find only 2 cm of crumple paper against only one of the 6 sides of the 2 TByte Drive,
and around the drive itself there was only 2 layers of Bubble wrapping, and 50% of the bubbles were popped and useless.

I would love to trust Amazon's choice of courier,
but instead I will have to settle for the results of whatever tests you suggest as relevant.

Hard Disk Summary
-------------------
Hard Disk Number . . . . . . . . . . . . . . . . : 0
Interface . . . . . . . . . . . . . . . . . . . : S-ATA Gen3, 6 Gbps
Disk Controller . . . . . . . . . . . . . . . . : Standard Dual Channel PCI IDE Controller (ATA) [VEN: 11AB, DEV: 6121]
Disk Location . . . . . . . . . . . . . . . . . : Channel 1, Target 0, Lun 0, Device: 0
Hard Disk Model ID . . . . . . . . . . . . . . . : WDC WD20EZRX-00D8PB0
Firmware Revision . . . . . . . . . . . . . . . : 80.00A80
Hard Disk Serial Number . . . . . . . . . . . . : WD-WCC4M2389833
Total Size . . . . . . . . . . . . . . . . . . . : 1907726 MB
Power State . . . . . . . . . . . . . . . . . . : Active
Current Temperature . . . . . . . . . . . . . . : 31 °C
Maximum Temperature (Ever Measured) . . . . . . : 33 °C, 08/06/2014 17:27:22
Minimum Temperature (Ever Measured) . . . . . . : 24 °C, 08/06/2014 08:19:57
Power On Time . . . . . . . . . . . . . . . . . : 0 days, 12 hours
Estimated Remaining Lifetime . . . . . . . . . . : more than 1000 days
Health . . . . . . . . . . . . . . . . . . . . . : #################### 100 % (Excellent)
Performance . . . . . . . . . . . . . . . . . . : #################### 100 % (Excellent)
Regards
Alan

Hi Alan,

I completely understand and agree - and I also really prefer extensive testing on drives I'd trust and prefer to use for many years.

As described at http://www.hdsentinel.com/faq.php#tests ( Hard disk health is low or recently changed or I just installed a new (used) hard disk. How can I perform a deep analysis? ), I always recommend

1) Disk -> Short self test
2) Disk -> Extended self test
3) Disk -> Surface test -> Read test
4) Disk -> Surface test -> Reinitialise disk surface

Because the last one (generally the most important to verify both reading, writing, data storage capabilities) is a destructive test, which would remove all partitions from the hard disk, there is no point in partitioning / formatting the drive before those tests.

After these tests - if both are passed and there are no problems reported - even a quick format is enough, as we previously verified that the complete hard disk is usable, all sectors can be read/written properly.


> I have not seen any "Conveyance" Test Option.

Not all drives support this test, so it may be not used. But the other tests are enough to verify the integrity of the sectors.

Many thanks.

It has passed these tests so far
1) Disk -> Short self test
2) Disk -> Extended self test
3) Disk -> Surface test -> Read test

I will now Reinitialize the Disk surface.

I am glad that I did not partition and format before testing.

Please note a BUG or inconvenience with the Surface test -> Read test :-

When I launched the read test the first time your main GUI was occupying most of the screen and could NOT be shifted out of the way whilst the Surface Test Pop-Up was on-screen.
Then I noticed the "Background" button and that hid the Test Pop-Up and allowed me to shift the main GUI so I could browse the Internet.
Some while later I noticed that amongst all the icons below the menus was one with a magnifying glass that identified itself as "Hard Disk Test",
and this said there was another hour to run, and brought the pop-up back out of the background and again on-screen.
I put the Test Pop-Up back into Background mode.
Two hours later there was no Test Pop-up, and the "Hard Disk Test" could not show it to me.
I cannot see any results from that test. That was almost 5 hours of wasted testing.

Today I repeated the test WITHOUT putting the pop-up into background,
and after 4 hours 50 minutes it now shows :-
Extended Self-Test
Successfuly Completed
Response 0x00
Elapsed Time 04:50:30

SUPPLEMENTARY QUESTION about Partition Alignment on this Advanced Format Disk :-
May I ignore this alignment requirement and be confident that it will all be correct if I use Windows 7 Disk Management to create the partitions.
Should I be apprehensive about some third party partition managers ?

Regards
Alan

> Please note a BUG or inconvenience with the Surface test -> Read test :-

I suspect you mean the Disk menu -> Extended self test. That can be sent to the background, as the Surface test -> Read test runs in a completely different window, which can be accessedd independently from the main window showing the hard disk status.

> When I launched the read test the first time your main GUI was occupying most of the
> screen and could NOT be shifted out of the way whilst the Surface Test Pop-Up was on-screen.
> Then I noticed the "Background" button

This is true for the Disk menu -> Extended self test (and Short self test).
Yes, then you can use the "Background" button, exactly as you wrote to send the test window to "background" and access the main window instead
Alternatively, you may use the Windows "Show Desktop" button (or switch to any other task from the task bar or by ALT+TAB) to hide it completely and access any other running task.

> Two hours later there was no Test Pop-up, and the "Hard Disk Test" could not show it to me.

Yes, when the test completes, there is absolutely no point to show the test window any more.

> I cannot see any results from that test. That was almost 5 hours of wasted testing.

The results is automatically saved and displayed on the Information page. See S.M.A.R.T. details section, which contains the last short / extended self test status, when it's started, how long it take and what was the result of the test. There is no wasted time of course

> SUPPLEMENTARY QUESTION about Partition Alignment on this Advanced Format Disk :-
> May I ignore this alignment requirement and be confident that it will all be correct if I use Windows 7 Disk Management to create the partitions.
> Should I be apprehensive about some third party partition managers ?

If you partition the drive under Win 7, then you can be sure that the partition is correctly aligned. There is no need to use any third party partition managers for that.
And you can very easily verify the alignment of the partitions in Hard Disk Sentinel: just click on the appropriate partition in the bottom left area in Hard Disk Sentinel and check the displayed details on the right.

Generally, if you see that the alignment is 4 KB or more (which is multiple of 4 KB) in Hard Disk Sentinel, then it is perfect.

Partition created under Win7. Note the alignment is 1 MB, which means that the partition starts at MB boundary.



Partition created under XP. Note the alignment is 512 bytes, which means that it would be not aligned to a 4 KB sector of the AF drive.

The results is automatically saved and displayed on the Information page. See S.M.A.R.T. details section, which contains the last short / extended self test status, when it's started, how long it take and what was the result of the test.

Many thanks, I see it now.

I am now 16 hours into "Reinitialize disk surface" of my new 2 TB drive,
with another day to go.
My family have complained about the computer noise last night.
If I put the computer to sleep for silence tonight,
will this process cleanly resume in the morning when I wake the computer and log in again,
or might it degrade one or more of the sectors being processed ?

I am using Windows 7 Ultimate,
and HDSentinel v.4.50.5 PRO seems to resume normal monitoring after I place the computer into SLEEP mode during my lunch break,
so I hope the initialization of a surface would also resume correctly.

Regards
Alan

I know that it can take really long time, especially on USB connection, I just finished testing of a 3 TB drive which took 21+ hours

Generally, there should be no problems if the computer went to sleep and then wake up, the test continues.
Just the curent block (which was under testing when the system went sleep) may show darker green color as the performance monitoring may reflect that it took more time to spin up and seek before continuing.

Anyway, you may simply stop the test and start it by Disk -> Surface test, but before starting it again, select the Configuration tab and specify that the test should start at block XXXX, where previously it stopped. For example, if 50% of the test completed, then you can specify that the new test should start at block 5000 - to test the remaining area.

Thanks, but I went to bed before that reply so I left the P.C whirring through the night.

The write test found no defects but the speed was diabolical

The READ test shown previously had a speed of 145 MB/s at the start end of the Disk,
and fell down to only 70 MB/s at the other end.
I see absolutely no point in WDC providing a SATA 3 interface rated at 6 Gb/s = 750 MB/s because the Disk is data density is inadequate for the task

The WRITE test was deplorable - a constant 17 MB/s from one end to the other as per sshot-60.png
Is it possible that the 64 MByte Cache withing the WDC is actually a Flash Drive with all the slowness of writing to a flash Drive ?

I have created a 100 GB partition and used Zentimo xStorage Manage to perform speed tests on both this 2 TB disk and the old Samsung 960 GB disk :-

WDC WD20EZRX-00D8PB0 ATA Device
V-100GB-2TB (V:), 97.7 GB "NTFS"
Type of file Speed of reading Speed of writing
Small files (32.0 KB): 21.10 MB/s 13.50 MB/s
Medium files (3.0 MB): 115.48 MB/s 53.90 MB/s
Large files (100.0 MB): 138.17 MB/s 55.15 MB/s

SAMSUNG HD103SJ ATA Device
C_SAM_System3 (Not mounted), 25.0 GB "NTFS"
Type of file Speed of reading Speed of writing
Small files (32.0 KB): 21.69 MB/s 19.39 MB/s
Medium files (3.0 MB): 104.02 MB/s 149.85 MB/s
Large files (100.0 MB): 125.76 MB/s 155.46 MB/s

As you can see above to 6 Gbps WDC is very slightly faster than the 3 Gbps Samsung when reading, but 3 times slower when writing.
Please advise what tests I should perform so you may give further advise upon what is going wrong and what I can do to remedy the situation.
Regards
Alan

> The READ test shown previously had a speed of 145 MB/s at the start end of the Disk,
> and fell down to only 70 MB/s at the other end.

Yes, this is fine: the "end" of the hard disks are much slower, yes, around 50% of the speed of the beginning due to the number of sectors in the outer tracks vs. inner tracks which can be read/written in a single rotation.

> I see absolutely no point in WDC providing a SATA 3 interface rated at 6 Gb/s = 750 MB/s because the Disk is data density is inadequate for the task

In theory, 6 Gb/s can perform 600 MB/s because this technology uses 10 bits (8 data bits + 2 for error correction).
Yes, the platters are slow, but the transfer between the internal cache memory and the computer can be fast. This is where the higher bandwidth can help.

The slower write speeds are really interesting. Maybe related to the current driver of the controller/motherboard, which works slower than expected.
Can you please use Report menu -> Send test report to developer option in Hard Disk Sentinel?
Maybe it can give some ideas, thoughts. Thanks!

I have sent the report.

Supplementary :-

1.
HWINFO includes this information for each of the Disks

[Transfer Modes]
Sectors Per Interrupt: Total: 16, Active: 0
Max. PIO Transfer Mode: 4
Multiword DMA Mode: Total: 2, Active: -
Singleword DMA Mode: Total: -, Active: -
Ultra-DMA Mode: Total: 6 (ATA-133), Active: 6 (ATA-133)
Max. Multiword DMA Transfer Rate: 16.7 MBytes/s
Max. PIO with IORDY Transfer Rate: 16.7 MBytes/s
Max. PIO w/o IORDY Transfer Rate: 16.7 MBytes/s
Transfer Width: Unknown
Native Command Queuing: Supported, Max. Depth: 32

I am now guessing that what I estimated as 17 MB/s was actually 16.7 MB/s in PIO mode.

SWI shows my drives include

Drive Drive Type Bus Type File System Total Size Used Space Free Space % Free
R:\ (WDC_Rescue) Local Disk SATA NTFS 3.1 742.1 (GB) 717.4 (GB) 24.7 (GB) 3 %
V:\ (V-100GB-2TB) Local Disk ATA NTFS 3.1 99999 (MB) 2480 (MB) 97519 (MB) 98 %

All my 960 GB Samsung Disk partitions are like R:\, connected to a 4 channel SATA interface
My 2000 GB WDC Disk is connected to a dual channel PCI Express interface and I observe it is shown as ATA rather than SATA.
I think it may be worth connecting to a proper SATA port.

CrystalDiskMark shows an appallingly slow Write when testing V:\, which is 100 GB occupying the first 5% of the new 2000 GB Disk

-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 © 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 146.163 MB/s
Sequential Write : 68.089 MB/s
Random Read 512KB : 49.389 MB/s
Random Write 512KB : 62.036 MB/s
Random Read 4KB (QD=1) : 0.592 MB/s [ 144.4 IOPS]
Random Write 4KB (QD=1) : 1.916 MB/s [ 467.8 IOPS]
Random Read 4KB (QD=32) : 0.667 MB/s [ 162.7 IOPS]
Random Write 4KB (QD=32) : 1.978 MB/s [ 483.0 IOPS]
Test : 500 MB [V: 2.5% (2.4/97.7 GB)] (x1)
Date : 2014/06/16 11:14:22
OS : Windows 7 Ultimate SP1 [6.1 Build 7601] (x64)
500 MB test on V: (2 TB Disk - slow write)

CrystalDiskMark shows the old 960 GB Samsung drive is more than twice as fast in the write test on the first partition

-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 © 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 146.942 MB/s
Sequential Write : 142.644 MB/s
Random Read 512KB : 44.533 MB/s
Random Write 512KB : 63.253 MB/s
Random Read 4KB (QD=1) : 0.538 MB/s [ 131.2 IOPS]
Random Write 4KB (QD=1) : 1.289 MB/s [ 314.7 IOPS]
Random Read 4KB (QD=32) : 1.147 MB/s [ 280.0 IOPS]
Random Write 4KB (QD=32) : 1.226 MB/s [ 299.2 IOPS]
Test : 500 MB [F: 77.8% (19.4/25.0 GB)] (x1)
Date : 2014/06/16 11:09:09
OS : Windows 7 Ultimate SP1 [6.1 Build 7601] (x64)
500 MB test on F:

Both Read and Write stay above 100 MB/s for all partitions within the first 70% of the Samsung Disk.

Regards
Alan

Thanks for the report and the information.

I suspect the problem is related to the disk controller where the hard disk is connected (or more precisely, its driver).

As you can see on the Information page in Hard Disk Sentinel:

for the 2TB WD hard disk:
Disk Controller . . . . . . . . . . . . . . . . : Standard Dual Channel PCI IDE Controller (ATA) [VEN: 11AB, DEV: 6121]

for the Samsung hard disk (and the other WD drive and your SSD):
Disk Controller . . . . . . . . . . . . . . . . : Standard AHCI 1.0 Serial ATA Controller (AHCI) [VEN: 1002, DEV: 4380]

The 2 TB drive is connected to a Marvell controller which has no drivers installed. Generally, the Standard Windows drivers are fine for some controllers (especially if the controller works in AHCI mode, like the AMD chipset controller where the other drives connected).
But ATA mode (and this Standard driver for that) may cause reduced performance with some controllers, as you can see.

I'd recommend to simply connect the 2TB hard disk also to the other controller on your motherboard, where the other drives connected. If there is no free SATA port there, you may simply try to swap with the 640 GB WD hard disk or the optical drive, at least to check the difference.
I suspect the write speeds will improve greatly then.
Alternatively, you may try a better driver for the Marvell controller - but it may not help, so I'd connect a device which may not require as high performance (the lowest capacity HDD).




[quote="alan-0000"]I have sent the report.

> I am now guessing that what I estimated as 17 MB/s was actually 16.7 MB/s in PIO mode.

No. If that would be the case, then the read speed would be also so slow.


> My 2000 GB WDC Disk is connected to a dual channel PCI Express interface and I observe it is shown as ATA rather than SATA.
> I think it may be worth connecting to a proper SATA port.

Yes, I agree: this would be the best.
That tool shows ATA as the controller is working with standard ATA driver. I suspect it may show SATA when the AHCI mode is used, but I'm not sure.