Hard Disk Monitoring

I have used 64 bit Windows 7 Ultimate for over 1 year and added the SSD 3 months ago
I installed your Trial HDSentinel Pro 4.0 on 23 May 2012.
I have configured Trial HDSentinel Pro to start with Windows and accepted the initial default of 5 Seconds.
I chose to set notifications on Disk #2 = OCZ SSD as icon + notifications
and Disc #0 and #1 as notifications only.

Until this morning I found
The first drive was always the 600 GB WDC Drive - non-bootable Secondary HDD
The second was always the 931 GB Samsung Drive - MBR bootable Primary HDD
The third was always the 56 GB OCZ Drive - MBR Bootable and Active System SSD

The above sequences were as shown by :-
Windows native Disk Management, which numbers them 0, 1, and 2,
and third party tools :-
HDSentinel :- which numbers them 0, 1, and 2
Minitool Partition Wizard :- which numbers them 1, 2, and 3
Macrium Reflect Image backup :- which numbers them 1, 2, and 3

MY PROBLEM :-
This morning OCZ and Samsung swapped Drive Numbers and places in the order of presentation by ALL four tools above ;
and instead of the Notification tray showing only the OCZ Sentinel Engine icon it shows only the Samsung Sentinel Engine icon.

Even though Windows changes the drive numbers HDSentinel still reports the Disk Locations as :-
600 GB WDC Drive, Channel 1, Target 0, Lun 0, Device: 0
931 GB Samsung Drive. Channel 2, Target 0, Lun 0, Device: 0
56 GB OCZ Drive Channel 0, Target 0, Lun 0, Device: 0

All I can think of as a significant change is that last night before shut-down I added to my Desktop Startup folder a link to my PaleMoon Browser
so that it would launch when I logged in, so Initial starting of Windows may take 2 seconds longer

Is the drive number allocated after I log into Windows and thus affected by a slower start-up ?

I am puzzled by the arbitrary nature that some tools number drives starting at 0, and others start at 1.

This erroneous change persisted after :-
a Windows Sleep and then Wake-up;
And a Windows Restart.
It was cured after I shut down the computer and then restarted.

I hope it does not go wrong again and would appreciate advice that might avoid future problems.

Regards
Alan

Thanks for your message.

In general, Windows associates the drive numbers for each hard disks during booting up the system.
The order is determined by various factors, for example the order of hard disk controller(s) (their position in the motherboard), the order of USB drives (if any) and so.

As you may see, all software follows this list. Some of them (for example Hard Disk Sentinel) uses the same numbering methods as you can see in the Windows Disk Management (starting from 0) for compatibility issues, while some software adds 1 (so starting disk numbers with 1 instead of 0) to help readability - but cause confusion (especially when having lots of hard disks with the same model).

I'm not sure the actual configuration, but according the information I suspect the drive(s) may experience a spin up issue: after the computer wakes up from sleep state, the drive(s) may not be able to spin up again in the order they previously detected by Windows. This is why the order is changed, as now the last drive may spin up sooner than the other.
This is why the restart of the system cured the problem: as when both hard disks are spinning, the OS finds them and assigns the correct drive numbers.

Personally I'd recommend power connections / cables. If the drives are external USB drives (especially if they're 2.5" drives) I'd recommend to use double Y USB cable to provide enough power for proper operation. Please check http://www.hdsentinel.com/hard_disk_case_spin_retry.php for more information about this topic.

Also you may use Report menu -> Send test report to developer option. This always helps to check the current status and may advise based on that.

Thanks for your reply.

All three drives are internal and connected by SATA leads.
Channel 0, Target 0, Lun 0, Device: 0 is my 55.9 GB SSD by OCZ, normally Disk no. 2
Channel 2, Target 0, Lun 0, Device: 0 is my 931 GB HDD by Samsung, normally Disk no. 1
Both the above have MBR and are bootable by my BIOS.
Channel 1, Target 0, Lun 0, Device: 0 is my 596 GB HDD by WDC, always Disc no. 0
This WDC HDD is unbootable by my BIOS because it is Basic GPT

Both my HDD have a Ready Time (Typical) quoted of 11 Seconds.
I have observed a 5 second delay between clicking a desktop shortcut and the launch of a *.BAT script held on the WDC HDD which Windows had put into power save mode.

The SSD does not need to "Spin Up" nor seem to quote a Ready Time.

I can understand WDC always winning the "Spin Up" race against Samsung,
but an SSD losing the race against both HDD is unexpected

I created a text report yesterday when everything is normal and have now attached it.
I wish I had thought of creating a report when the system was abnormal.

Regards
Alan

Dear Alan,

Thanks so much.
I thought that the drives are connected to different hard disk controller(s) and they ready-time after wake-up may be different, This could cause the issue.
But in the report I see that all drives are connected to "Standard AHCI 1.0 Serial ATA Controller (AHCI)".

I agree, interesting situation

Each drive has "Smart Details" which show several "self-tests" have never been started.

I have not yet run any tests because I do not know what they do or the effect on the SSD Life.

Would any of the tests give useful information upon this "interesting situation" ?
Are there any other diagnostic tests that you can recommend ?

I consider 64 bit Windows 7 Ultimate + SP1 to be totally defective in this situation.
Disc ID numbers should not get transposed willy nilly.

Please note that when the problem occurred I was immediately aware because I have not performed Temperature Calibration,
and the HDSentinel Engine for Drive #2 was showing 22 instead of 30 when booted up from cold.
This caused me to hover over the icon and see that #2 was now Samsung and not OCZ.

I would be quite distressed if a Windows startup SNAFU resulted in HDSentinel monitoring a non-vital downloads drive and ignoring precious archives or the system SSD

Suggestions :-
Base the HDD identification on the first partition letter or GUID instead of the variable Drive #X
and/or give a high priority alert if any particular non-removable Drive #number powers up with a change of first partition letter or GUID.

Regards
Alan

The tests are 100% safe for the stored data and they should not affect the life or the usability of the SSD, so you can any time use them.
However, in the current situation the issue may be more related to the communication between the devices (eg. the hard disk controller, the driver, the operating system) than the actual devices.
Personally I'd use these tests (on both the SSD and the hard disks) - just to confirm that there are no problems with them, but also I may check for an updated drivers for the chipset as that may have better power management functionality.

I can confirm the software uses the proper methods (identify drives by model ID + firmware version + sertial number) to assign the tray icons and identify the values. This is better than using GUID / partitions / letters as this is independent from the stored information (works even on unformatted drives or drives with unknown eg. Linux partitions).

Maybe you saw that after wake up, Hard Disk Sentinel may display the old temperatures detected, but as soon as notices that a wake-up event occured, it updated the temperature values to display the current, actual values. These should be lower than the previous values as the devices had time to cool down to the temperature of the room / office.

If possible, if you see it next time, please
- save a screenshot about the tray icons
- use Report menu -> Send test report to developer option
(this is better than the simple TXT report to investigate issues as the developer report contains "raw" information provided by the disk controller).

Thanks

I refrained from "send report" because I was uncertain whether that would fail due to my having not yet registered this trial Pro version.

I am close to restoring an image of C:\
If I register a license key now I assume I will need to re-register and possibly re-install when I put the system back to last month.
Will I have any hoops to jump through so I may re-use the same key on this computer ?

Please note that the unexpected low temperature alerted me to something strange.
Within a few minutes I had opened the HDS GUI and saw that in the left column
the Samsung HDD had become Disc 2 and and the OCZ SSD was Disc 1
and the Samsung and OCZ Health and Temp status had swapped positions.

Then I used in turn each of
Windows Native Disc Management and saw the graphic partition layouts also showed the same transposition of Disc number and height positions,
and the same again with third party tools Minitool Partition Wizard and Macrium Reflect Partition Image.

N.B. Samsung always starts at 22 degrees C and warms up to 30 quite quickly
The others start at 30 and OCZ stays at 30 but WDC climbs to 34 after an hour.

Regards
Alan

Thanks.
Of course you can use "Send test report" in any version.
This is designed exactly to check "interesting" situation and/or hardware compatibility issues. May be more frequently used in un-registered trial versions as then I can assist (for example recommending a driver update) to get all kind of hard disk / SSD information and fully utilise the registered version.

If you have registered version, you may use Configuration -> Update -> Backup configuration function.
This saves all settings, statistics, registration details, etc... to a single file. After re-install (or restore OS image) you can any time re-load this saved configuration file in Hard Disk Sentinel to continue using it, without the need of re-register the product.

Yes I see - of course the temperatures (and tray icons) follow the change of configuration. I completely agree that it is not expected that Windows changes the order after wake-up (I'd be also surprised to see) but as you can see in Disk Management, HDSentinel and in other tools, all should follow this change.

Ps. the Help -> Appendix -> temperature calibration would be interesting in this situation: Samsung drives usually report 7-8 Celsius cooler than other drives. So if all other drives report 30 Celsius immediately after startup (after enough time to cool down) but Samsung shows 22, the temperature sensor need to be adjusted (specifying +8 in the Offset column for attribute 194 Disk Temperature on the S.M.A.R.T. page of the Samsung drive) to reflect the _real_ temperature.

Thanks

I will get registered and do further testing and then return.

NB The OCZ SSD always reports 30 degree C.
You show this as being the minimum and the maximum and the average daily reading.
I just checked with Piriform's Speccy and this sees the same.
The other WDC HDD shows extremes of 23 to 39, and the Samsung HDD from 22 to 28

Is it expected that the OCZ-Vertex2 series will always say 30 degrees ?

Regards
Alan

Windows goofed again.
Disc numbers transposed.

OCZ SSD is normally Disc 2 , not Disc 1
SamSung is normally Disc 1, not Disc 2
Control Panel/.../Notification Tray should show Icon for OCZ SSD

Screen shot of HDSentinel and Notifications attached.
Report has been sent.

Regards
Alan

I observed this morning at about 11:00 that when a smart value is selected a history chart appears below.
I was horrified by the Raw Read Error Rate on the 7th June, a day when I never noticed anything wrong.

I have just looked again and suddenly I see disaster in the last few minutes
2,863,349 errors smart no 195
see sshot-33.gif
Then I registered on-line my license number and it immediately surged up to
2,928,925 errors.
see sshot-34.gif

The history charts for smart no 195 and no 1 are identical, apart from their titles.

On the 4th June Windows mixed up my Disc numbers but there were no errors
On 7th June there were 2,752,035 errors yet I was not aware of erroneous Disc Numbers - but I cannot claim to perform "due diligence".
Should I assume that uncorrected error count and Disc number errors are unrelated ?

How big is 2,928,925 errors ?
is it 2,928,925 * 2 bit cells = 732231.25 bytes = 178 off 4 kByte file clusters ?
Since the SSD firmware has detected the errors, does it re-read and deliver the correct information ?
Excuse my rising panic, but what happens if it cannot succeed in correcting the error ?

Regards
Alan

I am seriously concerned about the SSD errors. Within 2 minutes of my taking the first shot of 2,863,349 errors for no cause,
there was a 65576 increase - perhaps about 0xFFFF = 1 bank of a chip ?

I Put Windows into Sleep mode for a lunch break, and after hitting keyboard to wake windows,
and found all errors were zero, both Smart and HDS History Macrium Reflect created a FULL IMAGE of SSD (reading all used space apart from 8 GB of Pagefile.sys This instantly caused 27,???,??? errors
Updated Macrium Reflect and created an Incremental IMAGE of SSD (very little reading of only what had been updated)
Still at about 27,???,??? errors
Macrium Reflect created another FULL IMAGE of SSD Within about 3 minutes of a 27,000,000 error increase
there was a further 63150 increase - perhaps about 0xFFFF = 1 bank of a chip ?
I shut the PC down for a tea break after which I booted up through BIOS and into Windows.
This corrected the Disc Numbers as expected.
Unfortunately all errors were zero, both Smart and HDS History Macrium Reflect created another FULL IMAGE of SSD Within 16 minutes of 26,824,687 errors there was a further 160,999 = 0x274E7 increase.
After 2 hours the increase rose to 1210378 = 0x12780A

I am concerned that the SSD has only 19 GB of used space, of which 8 GB is PageFile.sys,
and simply reading the other 11 GB of used space will typically cause 28,000,000 errors where an error could be 1 byte, 4 kByte, or 500 kByte.
I am also concerned that 0xFFFF to 0x27FFF subsequent increase denotes that 27,000,000 errors when reading 11 GB caused the firmware to "pension off" the weak data blocks that had errors,
and available date had to be copied by the firmware to fresh blocks and the internal copy also caused errors.

I would appreciate advice upon what HDS tests to run and whether these errors are to be expected or if this SSD needs urgent replacement.

Please note that Error counts in the Smart Data from the SSD are reset when the PC is either rebooted or simply put to Sleep and then woken,
and unfortunately HDS only shows the latest current SSD data for "today",
so major errors can be seen and shown, but after a reboot or sleep they will be forgotten unless, perhaps,
the errors existed until shutdown at the end of one day then perhaps HDS will have the end of day errors captured in History.
I now fear that the 11 days in history with an error count of zero may have had 9 digit errors each morning before a lunchtime sleep.

I have already posted sshot-33.gif and shot-34.gif.
The others are now attached

Regards
Alan

only 3 attachment allowed, so here is 38 and 39 and a zip of 40, 41,42,43

Problems with HDS :-

HDS starts with Windows.
Daily Average, Daily Maximum, and Daily Minimum temperature readings are show for my SSD and both HDD.
The numbers are believable apart from the SSD which never deviates from 30 degrees C.
The history chart shows these numbers for each and every day between, and including, 2nd through to 11th June

SMART DATA IS LOST.
When I select ANY SMART DATA such as Disc Temperature I can only see values for
02/06/2012, 03/06/2012, 05/06/2012, 06/06/2012, 08/06/2012, 09/06/2012, 11/06/2012
No data nor date itself is shown for 04//06/2012, 07/06/2012, 10/06/2012
Obviously relevant smart data was available on these "missing days" or no Daily Average would be shown.

You will see from yesterday's post there were 2,752,035 errors on 07/06/2012, and over 20,000,000 errors on 10/06/2012.
Today no errors are shown, as attached.

I also have very grave concerns that all errors are totally forgotten if the computer sleeps or shuts down and is re-started the same day.
It is possible that errors at the end of the day will be remembered the day after next, but I will have to wait till tomorrow to see it it happens.

Regards
Alan

Sorry for the confusion.
I can confirm that the SSD temperature is not valid if it is always reported as 30 Celsius as it should change during use of the device.
Most SSDs absolutely have no temperature sensors, reporting no value at all.
Some manufacturers are "smart" enough to provide a such fake (invalid) value. This is the case now: as you can see, your SSD always report 30 Celsius regardless of the actual temperature or conditions.

So please check and compare the temperatures reported by the hard disks only. If they are similar (1 Celsius difference is possible of course) and they start to climb up from the office/room temperature, then the sensors should be fine.

alan-0000 wrote:Is it expected that the OCZ-Vertex2 series will always say 30 degrees ?

Thanks for the report and the screenshot.

Yes, I see the modified order of the drives.
However as I see, Hard Disk Sentinel correctly followed the new order and reported the correct values for both devices (30 Celsius (not real) for OCZ SSD, and 25 + 27 Celsius for the hard disks).
Previously I thought the issue was that different temperature was reported (eg. the temperature of the hard disk for the SSD) which is not really possible - but now I see that this is not the case, the temperatures follow the drives as should.

I can confirm that this high value is not the count of errors.
These are the total read operations performed by the drive, not the errors
(that's why the attribute is "raw read error rate" instead of "raw read error count").

When this attribute is selected, please check the "Flags" under the S.M.A.R.T. attributes,
displayed next to the graph of the selected attribute. The term "event count" is not
displayed - as this attribute does not count the number of statistical or error events
(in contrast for example attribute 5, 10 and so),
just reflecting the rate of errors compared to all operations. As you can see, "Error rate"
is displayed in this "Flags" section.

It is normal if you have increasing values (even millions) and/or sudden drops
for Raw read error rate and Seek error rate for this drive, please do not worry about them.
(this is described in the Help of Hard Disk Sentinel).

The change of these attributes alone does indicate disk problems
(except when they have really wrong values, then HD Sentinel would immediately
alert with the problem and significant drop in the health value).

If you are interested, please check
viewtopic.php?f=32&t=735&start=0&st=0&sk=t&sd=a
topic as this issue is discussed there.

Excuse me but as for the Seagate hard disk, these numbers do not indicate errors.

As you can see in the text description of Hard Disk Sentinel:

" The status of the solid state disk is PERFECT. Problematic or weak sectors were not found. "

While this showed, there are no problems and you do not need to worry.

The increase of these values are normal during read operation. This is why
- you see constantly increasing values especailly when you perform many read operations (for example disk imaging)
- you see clearing of the values upon power loss / restart as the SSD clears this internal counter

If you feel something strange and you prefer intensive testing, you can use the tests from Disk menu.
The short self test, extended self test, Surface test -> Read test functions can be used now (these are not destructive tests) and they will reveal if there are any problems found.

alan-0000 wrote:Daily Average, Daily Maximum, and Daily Minimum temperature readings

Yes of course - these values show the statistics based on the measurements performed during active periods.
The 30 Celsius is reported by the SSD. I'm afraid it is not correct - but this is not really related to Hard Disk Sentinel.

alan-0000 wrote:SMART DATA IS LOST.

I can confirm that no information / history lost and yes, the daily values show the _last_ information recorded on a particular day.
So if you see milliöns in the raw values increase but then the system restarted (and the behaviour of the SSD is to clear the particular statistical counter which does not really reflect error) then yes, I can confirm that the value 0000 is recorded and then re-displayed.

As you may see, some intermediate days are not displayed - but these values are not lost, just the graph is automatically adusted to show the trend of change. Look attribute #9, #12, #241, #242 and so - as they confirm that the values are not lost: they constantly increase and never decrease back to zero.

alan-0000 wrote:I also have very grave concerns that all errors are totally forgotten if the computer sleeps or shuts down and is re-started the same.
It is possible that errors at the end of the day will be remembered the day after next, but I will have to wait till tomorrow to see it it happens.day.

Of course no errors lost / forgotten. Exactly as you wrote, the last value measured is saved for each day.
I can confirm your SSD clears some statistical values upon power on/off cycles - but this is not related to Hard Disk Sentinel but designed by the manufacturer to work this way. Excuse me if this behaviour of the SSD caused confusion.

Please check the values for hard disks also (especially the constantly increasing values) to confirm that the values are properly saved and then re-displayed on the statistics.

hdsentinel wrote:I can confirm that this high value is not the count of errors.
These are the total read operations performed by the drive, not the errors
(that's why the attribute is "raw read error rate" instead of "raw read error count").

viewtopic.php?f=32&t=735&start=0&st=0&sk=t&sd=a
topic as this issue is discussed there.

Thanks for all the answers and for the link.

I was unhappy that often the SSD was in use but attributes #1 and #195 remained at zero,
and then whilst my use remained the same both #1 and #195 would jump, e.g. to 2,863,349.
The fact that they stay zero during normal use suggested a serious error when the value jumped to a high level.
#1 and #195 were always the same as one-another.
I can live with a Raw error if it can be ECC corrected.

The Flags for #195 state
Event Count, Error-Rate, Performance.
Does this mean that the value corresponds to whatever OCZ says that it means and not what I read ?

Regards
Alan