Hard Disk Monitoring

Not sure where to post this, since it is mostly feedback on HDSentinel (HDS) itself, with some feature requests added.

I only bought HDSentinel Pro a few days ago, and have spent the past couple of days tweaking settings (custom temp limits and much else to get alerts/warnings at the right level).

So far, I must say that HDS is a dream come true in helping to manage my system. With my workstation pushing 30+ drives under management, watching them all manually would be a pure nightmare.
With 17 drives running inside the chassis, and the rest of the 30+ drives in 2 JBODs and on USB connections, keeping track of their aging and failures without automation would be impossible!

As initially (too many) alerts came through in both alerts and email, it was initially a pain in the backside. But as I tweaked it all, it has been VERY useful!!

On the positive side:

1. HDS warnings about overheating drives made me go investigate one of the JBODs, and it turned out the 140 mm fan supposed to keep the drives cool had seized up. Not spinning.
So off to replace that fan with a spare to get some airflow again. No more HD alerts..

2. 3 drives turned out to be aging out with sector failures and relocations.. So off ordering replacements and moving data to new drives. Then these drives ended up in the trash..

3. A fourth drive had some pending/weak sectors but was saved by running a Surface initialization that took it back up to 95% health. Good enough to use it for at least temporary data moves and still be safe, unless HDS screams again.

4. 13 of my internal (chassis) drives are running behind an Adaptec controller, some of them in mirrors which presents them in Windows diskMgmt as single drives. The rest are passthrough and seen/enumerated directly in Windows. HDS sees and monitors ALL the drives individually, including the ones in Adaptec mirrors.. GREAT, GREAT! But see c) below..

4. Since I hate wearing on the $400 Nvme M.2 system drive system drive by blasting it with unnecessary writes, I had already moved TEMP files, and many app caching/logging of to a different drive.
Since HDS constantly writes to it's database and stats logging directory, it's default location on the system drive was not making me happy.
But I was VERY happy to see that HDS config supported directly moving the settings and stats folder off to a better disk.. Good stuff... Worked like a charm.

5. Custom thresholds for individual disks.. Good.. With different locations, and between 2 Nvme M.2, 3 SATA SSD, and a multitude of spinner drives, having them all set to the same temp thresholds would not be useful.

Overall VERY happy with HDSentinel.


Some things I wish were different or could be added...

a) A way to enter a user comment into the individual drive logs to consolidate notes in one place. For example, in my case the temperatures caused by the failed JBOD fan. It would be good if the log could have a user comment entered on that day/time to remind me later why all the alerts happened.

b) With many drives scattered on multiple controllers, the drive selection in the main window is a pain in the ***.
You get an alert, go to find that drive in HDS to look at the details and have to scroll slooowly through this long list, trying to recognize the drive. Aargh...
It would be good if there was selection filter(s) on top of the drive list to limit how many are shown (like filter down by controller, showing only disks on that controller). Maybe other filters.

c) On a simpler system, Windows Diskmgmt and HDS disk numbers would match up. But with Adaptec hidden mirrors, the disk numbers get skewed off by the number of "hidden to Windows" mirrors or RAID setups.
After meeting the first mirror the numbers no longer match up. Not a big problem, but it makes for a nervous guy when I get alerted to a bad drive (by HDS number) and then want to run a destructive surface scan.
HDS might say it is disk #26, and to make sure I destroy the right file-system I also look in DiskMgmt, but then have to watch out because in Windows it might be listed as disk #24. So also in Windows Event log, where logged events might say that disk #24 is having serious problems, while I have to watch in HDS, where the drive # might be 26 (or some other number). For destructive scans, one had better double, and triple check that the scan is run on the right drive.
Not sure if the Adaptec enumeration of drives allow you to match up which drives belong together in for example a mirror. If it does, it would be nice if in HDS such drives could be listed as 7A and 7B, instead of 7 and 8, which skews the numbering so it does not match with Windows. If drives belonging together (hidden from Windows) were listed with a letter suffix, the numbers between HDS and Windows would still match, instead of being offset.

d) It would be convenient if the settings includes a way to "turn off" the volume/filesystem listing showing space usage below the drive list on the main window.
For two reasons.
First, I do not monitor space at all with HDS. Useless info. That space could list disks instead.
Second, it only shows drives that has a letter assigned.. All my backup and media drives have no drive letters at all. They are mounted on directory paths. Only a select few has a drive letter (to not completely clutter up Explorer).
For example, all my backup disks are mounted on named directories under C:\BackupDisks, media drives under a different directory. Allowing be to see the disks's location in various bays should I need t find it.
None of these directory mounted drives show in the Usage section on the main screen anyway.

But all in all a very positive experience with HD Sentinel. Thank you...

Add-on to c) above about drive numbers.

HDS seems a little schizophrenic or confused about it's use of drive numbers.
Did another scan on a drive last night. Then suddenly noticed, that the Surface Scan window shows a different disk number than HSD itself.
As in, I am on disk 16 in HDS main window, then select to run a Surface test (Initialize in this case).
But when the Surface test window shows up, it's title bar claims to be doing the massive destruction on disk 14 (in my case the Windows disk #)..
Enough to induce a moment of panic to validate that it is in fact destroying the right drive; the one I had selected.

As mentioned in the original posting, the skew between Windows disk number (#14) and HDS number (#16) is caused by 2 mirrored sets that HDS discovered before this drive.
Each bumping HDS drive numbers by +1.

Thanks for your message and your feedback, good to hear that generally Hard Disk Sentinel work as should and designed

I'm sending the anwers to your tips and suggestions:

a) about user comments: the Overview page has a line to specify any comment for any disk drive. This can be used to specify a comment, eg. installation date, maintenance date (eg. fan cleaning/replacement), warranty expiration day or anything else. Not sure, but may help (?)

b) thanks for the tip about filters, will consider that. While it is not yet possible, if you right click on the list of physical disk drives, you can
- select to have smaller buttons
- hide (uncheck) any disk drive (designed to allow hiding pendrives/memory card readers, but you can hide other disks - however I prefer them to be listed to prevent confusion)
These may help to show more important details there and focus on the problematic ones.

About alerts: you can consider the options at Configuration -> Message settings page:
"Include problem overview in alerts" - to show a quick overview of problematic disks on the top of the alerts
OR
"Alert about problematic drive(s) only" - to receive alert specifically only about disk(s) triggering an alert

This may help to quickly find out which drive is affected.


c) Thanks, but of course Hard Disk Sentinel is absolutely NOT "schizophrenic or confused" about the disk drive numbers

Sorry for the possible confusion, but yes, exactly as you noticed things designed as you wrote:
- generally Hard Disk Sentinel follows the disk numbers you can see in Windows Disk Management exactly to allow identification of drives easier.
- there is an exception: when a RAID array found. Then (exactly as you can see) instead of showing a single virtual array (which appears in the Windows Disk Management) the RAID members (2 or more) are displayed and this causes that further disk numbers in the main window are adjusted according that.

Thanks for the tip with the 7A or 7B or similar - but that would cause even more confusion.... Plus what happens with a RAID array of over 26 disks

BUT: in the Disk Surface window always the numbers appear in Windows Disk Management displayed.
This is normal and expected because (as the purpose of RAID is exactly to prevent disks to be read/written independently) the disks of the array need to be managed (tested) together.
This is why the Disk Surface window (when you select the array there) shows two physical disk drives (if two drives form the array) exactly to confirm that BOTH of those drives will be affected.
And this is why the test window shows confirmations, just to allow to triple-verify if the proper disk is about to be tested/overwritten (by disk serial number or so).

Maybe to make things more clear, the Information page can show an additional entry on the top, eg. "Windows Disk Management Disk Index" or similar (if this is different than the drive number you can see in the upper left area, so if a RAID array found) as that can also help the properly identify the drive.


d) While that area can't be hidden, you can click on the slider above the first logical drive (separating the physical disk selection - and logical disk selection areas) and drag to bottom to minimise the bottom part and increase the top part. Can help, especially with the options described in b)

Thanks for your kind words and the tips, suggestions, they'll be considered in future development to make things more clear and usable

On a) User comments. Yes, I use that comment field to add extra info on certain disks, which then show on the disk list. Very useful, but unfortunately would not help, as it is not showing paired with the log-entries (logged with the event they explain), and is too limited.

On b) The filtering of disks. I tried to change the button "sizes", but that really only works by limiting the information shown. I prefer the large buttons to see as much info as possible.

I thought since, that a much easier and probably more useful way would be to add a simple search/filter text-field up top, where one can type a text string. Also likely the easiest to implement.
The list of disks being enumerated could then be limited to showing disks where one or more of the text info fields (model, serial, disk-label, mount-point, comment string) on the disks match that filter string, if not blank..
Part of the serial, limit to certain models (type in "WD" and all disks with WD in their model would show), type in part of the disk-label, ...
Whatever identifying info makes sense to the user.

If that text/search field supported entering a standard regular expression, even better. For example, all my disk-labels include a physical location identifier (so I can find the drive in case of trouble).
Regular expression support would allow search strings like 'B[0-9]$' for my chassis right-side 10 bay drives, 'W5$' limit to the disk in wall-carrier 5, 'R[0-9]$' to limit the list to all drives in one of my JBODs, '^WD.*' to restrict the list to WD models, .

You get the point. A fairly simplistic limitation to implement, where as you enumerate the disks in the list, each drive is displayed ONLY if one of it's text fields match the search string.


On Alerts. Yes, I use that selection already to select content of email alerts. I had to check the "Alert about problem disks only", as the default of showing all disks make the alert emails totally impossible in size and irrelevant. Get an alert caused by one drive, and the email contains info on 30+ drives to scan through to find out what the problem is. Not possible..

I can see your point on the disk numbering, although there is really not a normal RAID config I know of that will hide anywhere close to 26 drives as a combined cluster of disks, all grouped together behind a single Windows disk number?
A RAID cabinet can surely contain many drives, some of them combined in various RAID configs, with each RAID grouping hiding behind a single Windows Disk#, but I still thing alphabetic suffixes would work. No individual Windows disk # would ever hide more than 26 alpha-letter disks behind it.
Now that I know the numbers are skewed, that initial panic feeling is less. Maybe a slight change in the title bar would help, also showing the number(s) of the HDS disks involved would be useful. (HDS#/Win#)

Pulling down the slider/separator to "hide" or limit the volume/space section and see more physical disks. Yes, perfect.. Why did I not notice to do just that.. Works great.


Add-on about the email alert subject lines..
With the number of alerts I have seen now, I have learned that it would be very nice if the subject lines actually contained some useful identifying info.
All my subject lines are essentially the exact same, no indication of the drive causing the trouble, which forces you to open each email to see if it is merely a repeat alert or a new drive overheating.

For example, all my email subjects are essentially like "Hard Disk Sentinel: Overheat (ATHENA) [69 %]".

a) "(ATHENA)" is the same in all emails (since only one system is alerting). Maybe an option to turn off node-name to give more space for other info, if node name is not needed because there are multiple systems sending emails?

b) I realize that the "69%" is an indication of the lowest health drive overall, completely unrelated to the drive doing the alerting.
But that is fairly useless. I my case, the "69%" stems from an 8-year old SSD drive, still attached.
HDS gives it 69% only because of it's wear-leveling count, but Samsung Magician still claims the drive is "Good".
I do not let it carry important data anymore, but use it only for temporary, fast data moves, but reminding me about the 69% on every email about other drives seems less useful.
Better to just remind me about that drive when it has it's own problems or the health deteriorates further.
If the email subject line contained useful, identifying info about the drive doing the alerting would be better. Disk #, Model/Serial.

Like for overheat emails maybe change to subjects like:

"HDS: Overheat (Disk: #27: WDC WD1001FALS-00J7B1)"

Or here is another alert I got:

Hard Disk Sentinel: Critical value degradation (ATHENA) [70 %]

The actual issue was that a spinner drive had seen an uptick in bad sectors.
A very useful alert as such, once you read the email body, but there is no identification in the subject that relates to the actual drive in question.
(It is still ATHENA, and the 70% refers to that same old SSD drive that now later shows as "69%". The actual drive generating the alert is in 100% health.)
Also mystified that in the list of alert emails I looked at, there is no actual alert that the old SSD suddenly got downgraded from 70% to 69%.. That would have been a useful "degradation" alert.

But as I mentioned earlier, I really LOVE HDS.. VERY USEFUL. Great Job.
My wish-list is more from a developers perspective, that "Had I developed this software myself, I would like to have done this"...
Things that could make it even more useful..
So don't take it as negative feedback.. It is not meant that way. More as potentially useful changes that could make an already fantastic piece of software even more fantastic.

I realized a couple of items that might be useful monitoring items, although you might consider it a little out of scope for HD Sentinel.

I have had disks simply "falling out" (disappearing from Windows). Could be a JBOD powered off when rebooting, or it could be a cable not right/not seated correctly.
Windows will simply not detect them and one or more drive-letters silently disappear, or the directory it is mounted on is simply not mounted (turns into a normal directory), which is even less visible to the user until something crashes.
In HD Sentinel they simply no longer show up in your list. Again, silently.

With the default being exactly as it is, maybe in the Configuration -> Hard disk section, if one of the options could be a checkbox to mark the disk "permanent", which means that the disk's info should still be shown in the main screen's disk list, but maybe colored RED, as "missing". Plus an alert and email going out that this "Permanent" drive has gone missing from the drives tracked. With many drives, sometimes I don't notice that a drive fell out until something goes wrong. Backup failing, script crashing, ...

In that vein, a similar potential check to turn on could be a check that the file-systems are still alive. (Default still being that this check is off.)
Could be as simple as HDS putting a hidden, empty file in the root directory and each disk-check then simply checking if that file(path) is accessible.
Yes, that check should be off by default, as otherwise for people that let drives go to sleep, the drive would be kept alive forever. But all my drives spin 100% of the time. None are set up to spin down. (It is simply a bad idea.)
Why this check? I have had a case where a drive "vanished" and simply were not mounted on it's directory under my C:\0-BackupDisks. The next time something starts up, the path is followed, directories are created and the copy or backup simply proceeds to dump all it's data into the actual C:\ drive (where the mountpoint would have been). As a consequence filling up the C: drive 100%, since it simply do not have the space of the otherwise mounted drive.
I have even had disks that, maybe because of boot timing, suddenly Windows will see as "RAW". They show in diskmgmt as "RAW" (no partition/filesystems), while in reality being fine. A subsequent scan or reboot, reloading their info, the partitions/filesystems come back as just fine. Nothing wrong.
But a "file check" could alert the user if a file-system's drive's actual functionality is still available.

Since each drive could potentially contain more than one filesystem/volume, actually this last type of check probably belongs in the "Volume" section, if that section monitored/showed filesystems without a Drive letter (directory mounted only).
Which it really, really should.

Just some potentials to think about to increase the monitoring value of HDS..