Drive goes offline but HDS reports no errors

[mr-b]

Hi

I've found that one of my drives intermittently goes offline (it disappears from Explorer and Disk Management) and I have to go into Disk Management to rescan the disks and then it comes back. Event Viewer reports lots of these errors:
"Event 51, Disk - An error was detected on device \Device\Harddisk3\DR3 during a paging operation."

However HDS reports no trouble with this drive and reports its health as 100%!
Why does it not detect this error?

[hdsentinel]

Usually if hard disk intermittently goes offline, then the issue is more related to the operating environment, not the hard disk itself.
Cables, connections, insufficient power source are the most frequent causes of such situation.

If the drive is re-connected, then its status appears again - and if it's 100%, this may even confirm the above.

Hard Disk Sentinel (standard and Pro) shows the status of connected hard disks (and can display the last known status of offline drives, when they're not connected).
They issue no warning if the hard disk is disconnected or removed from the system. This is completely normal, as this function is available in the Enterprise version only at the moment (but because of requests, this may be available in a later Pro version as well).

Personally I'd recommend to
- check cables (both SATA data and power cables) and their connections.
- maybe replace SATA cable if you feel it does not provide good connection (not stable on the motherboard/controller or the drive)
- try a different power connector (if there is available on the power supply)
- avoid using power splitters / 4-pin -> SATA adapters

and then use the tests (as described at http://www.hdsentinel.com/faq.php#tests ) to verify the situation, confirm if the drive is really perfect - or reveal any kind of issues which are in the background.

[mr-b]

Tx for the info.

I can appreciate that there is a subtle difference between the health of a disk vs the availability of a (logical) drive.
However surely HDS can detect whether a disk has been ejected i.e. on purpose, rather than disconnected? Also I'd have thought i.e. the presence of Disk errors in Event Viewer should be warning that something is amiss?

Maybe this leaves open the door for a future value-added app "Storage Sentinel"?

I've reseated the SATA cables before but will try swapping them and the power cables over with another drive to see if the problem moves with the cables or stays with the drive. The issue is very intermittent though so it's hard to diagnose. Thanks again.

[hdsentinel]

Generally it is not possible to determine if the drive removed on purpose or disconnected - maybe except by the event viewer entries (but it may not happen if the drive was not acually performing a disk operation at the time of the removal).
Anyway, yes, exactly as you wrote, researching the possibilities and the future versions can issue even different kind of alerts, for example upon detection of such situtation.

Yes, exactly because it is hard to diagnsoe and re-produce, if you have the opportunity, please use
Report menu -> Send test report to developer option now and then later any time if you encounter similar.
This would help to check the situation, examine possible traces and compare the S.M.A.R.T. information reported (and its change related to the issue).
This would help development for this specific situation as well.
Thanks!

[mr-b]

Tx - have sent dev report as requested.

The functionality that I was specifically looking from HDS was the detection of these disconnections, since other than noticing the drive disappearance in Explorer, or if an app that writes to teh drive throws an error, there is no alert in Windows at all, just the Disk error in the Event Log. Maybe HDS could monitor the Event log for certain classes of Disk events in future?

[hdsentinel]

> I was specifically looking from HDS was the detection of these disconnections

This is available in Hard Disk Sentinel Enterprise version.
That checks and keeps log (and can issue alert upon) device connection and removal.

> there is no alert in Windows at all, just the Disk error in the Event Log.
> Maybe HDS could monitor the Event log for certain classes of Disk events in future?

Yes it can and it is planned.
However, it may not really accurate becuase of the following reasons:
- Windows can log such events even with perfect hard disks when the logical drive (partition) reports an error with the file system.
This may happen even with perfect drives.
- There are no such events logged if the drive removed before booting. For example if the drive was present on shutdown / restart but disappear before booting the OS.

Because of the above, this method not yet added, but it is possible that it will be in the future.

Monitoring drive connection / removal (in general) is now available in Hard Disk Sentinel Enterprise version - and (as I wrote, because of the requests) will be also in a future Pro version.

[mr-b]

Hi

I found that my machine was behaving oddly at/after boot - very slow response, some program execution errors etc. I found lots of these errors in the Event Log:

ATAPI Event ID: 11
The driver detected a controller error on \Device\Ide\IdePort2.

After some googling, it was suggested to try a HD Tune and this threw up SMART ID C7 Interface CRC Error Count, and it actually suggested checking the cabling. I replaced the SATA cable and the errors went away.
Are there any plans to include this type of error in HDS?

[hdsentinel]

This sounds more than interesting.

I can confirm that this is already included in Hard Disk Seninel for very-very long time: Hard Disk Seninel pays extra attention on such issues and report them in different ways to increase attention.

Hard Disk Sentinel not only shows the data communication issues ( see http://www.hdsentinel.com/hard_disk_cas ... _error.php ) but in the text description shows warning:

Problems occurred between the communication of the disk and the host 122 times.

and depending on the actual situation (if the situation is severe and stability problems may occur, for example on IDE/PATA hard disks):

In case of a sudden crash or reboot it is recommended to try a different, short data cable (avoid round cables, use 80 wire standard cables instead).


Plus in the Log page it shows the increased number of such data communication issues. Plus the above link on www.hdsentinel.com website shows even more details, information about how to diagnose, reveal and avoid such problems (not only by Hard Disk Sentinel, but also in general).

So it seems "googling" was not really required at all and especially lead to wrong direction, as this is included in Hard Disk Sentinel for years.

Please use Report menu -> Send test report to developer option.
This way it is possible to check the actual situation, the hard disk and the hard disk controller (motherboard / driver) combination, which may require special attention and then of course if modification required, it will be done as soon as possible.

Thanks for your attention - and I'm so curious to see the results and investigate.

[mr-b]

Actually now you mention the Log tab, I'd never looked at that (as the Overview page never mentioned anything).
The log page does indeed show lots of "Event #199 Ultra ATA CRC Error Count" increasing from n to n+ something.
Also on the SMART tab, that 199 value has Threshold of 0, Value of 80 and Worst = 80, Status = OK (Always passing).

So the CRC error was detected, but because the only thing mentioned in the Overview page was the 98% health determined by the SMART #177 Wear Levelling Count, and "No Actions Needed", so I just didn't think to dig any deeper.
What determines that a logged error makes it onto the Overview page?

[hdsentinel]

Thanks, now I see this is an SSD. Until now, you did not mentioned that, but I'm not surprised.

While attribute #199 Ultra ATA CRC Error Count is more serious on hard disk drives (and many SSDs), on some SSDs depending on the model / firmware, this may work slightly differently.
This is why Hard Disk Sentinel (while records the change on the Log tab) may not show on the Overview page for this model.
(also it is possible to configure alert to be issued on increased number, to be notified about it, please see Configuration -> Alerts -> When a new log entry is added).

Personally this is why I asked for a developer report with Report menu -> Send test report to developer option.
This way it is possible to check the actual SSD model ID and firmware version - and make Hard Disk Sentinel to pay extra attention and report issue on the overview page for this model as well (just like for hard disks and other SSDs).

> Also on the SMART tab, that 199 value has Threshold of 0, Value of 80 and Worst = 80, Status = OK (Always passing).

This is absolutely irrelevant.
This only means that regardless of the actual error count, it would never indicate failure / S.M.A.R.T. error when you start the computer (even if the BIOS S.M.A.R.T. checking enabled) because the "Value" (which decreases upon problems) can never drop below the "Threshold" (which is zero). This is why the "Always passing" displayed. Also it means that this regardless of the status of this attribute, we can't ask for warranty replacement.
(but this is normal - as the problems are not really related to the device, but the cables/connections).

The important is the Data field, marked with bold font. This contains the actual error count.


> What determines that a logged error makes it onto the Overview page?

The current device type (hard disk / SSD), model and firmware version. These together determine how the different self-monitoring S.M.A.R.T. attributes should be used, which are critical, which may indicate problems and which should be displayed on the Overview page.

[hdsentinel]

Sorry for asking, but did not really receive report about the drive, and would love to see its status and improve detection and reporting of problems for this particular device.

Is it possible to send a report - or at least some images/screenshots about the Information page (to show the disk model ID and firmware version) and the S.M.A.R.T. page to info @ hdsentinel . com ?

This would help to improve things and to show possible issues long before they can cause troubles.

Thanks for your help and time - and for increasing attention for this situation.