Jump to content
  • 0

a tip for fellow newbies


ikon

Question

I've just started using Scanner on two 5-bay eSATA/USB3 drive enclosures I inherited. They're great, but it can be a little hard to match up the drive model numbers displayed by Scanner with the drive letters and names in Windows Explorer. Fortunately, I happened to notice that it's possible to find out letter used for a particular drive by checking the Drive Details for that drive (the drive letter is shown a few lines down from the top of the window). Then, by checking the drive's label in Windows Explorer, it's possible to go into Drive Settings in Scanner and change the drive's name. I add the drive's label and drive letter just before the model number that's the default name Scanner assigns to each drive. So, if I have 4 drives that are part of my MainStorage pool, I might change the names in Disk Settings to something like:

MainStorage E: (WDC WD10EAVS-00D7B1)

MainStorage F: (WDC WD10EAVS-00D7B1)

MainStorage G: (WDC WD10EADS-00L5B1)

MainStorage H: (WDC WD10EADS-00L5B1)

 

This helps keep all the MainStorage drives together in the list, sorted by drive letter. I guess my next chore is to put a physical label on each drive, so I can quickly pick out one when needed.

 

ps. I'm so impressed with Scanner I just bought a 2nd copy. Now I have it running on my server and desktop (does it ever run fast on my desktop's SSD boot drive B))  Oh, and the 2nd copy was HALF PRICE! Awesome! 

Link to comment
Share on other sites

15 answers to this question

Recommended Posts

  • 0

Another trick?  If the enclosure has activity LEDs for each drive, you can use the "burst" test in StableBit Scanner.  This will create a bunch of activity on the drive, keeping it active.   You can use that to identify the drive, as well. 

 

 And yup, additional licenses are cheap! 

Link to comment
Share on other sites

  • 0

Thanks for that tip. My two Lian-Li EX-503 enclosures do indeed have activity LEDs. I tried using the Ping test to identify a drive, but it seemed to do nothing at all.

 

My one disappointment with Scanner so far is that it won't lock out bad sectors. I have to admit, I find that very baffling. It seems like Scanner is already doing all the heavy lifting in finding and identifying bad sectors (and relocating recovered files?). Is it really that hard to lock them out?

 

If it did that, my feeling, from a newbie point of view, is that Scanner could become much more usable, and popular, with the non-techie crowd. Imagine being able to tell people, "Hey, we've got this awesome little program that will sit in the backgound on your computers and keep your hard drive(s) in good shape. If it ever finds a problem it can't fix, it will notify you by email, so you can take action before you lose all your data". Basically, it would be like putting disk/file checking on auto pilot: non-techies LOVE things on auto pilot.

 

By the way, if Scanner won't lock out bad sectors, how is someone supposed to do it? You mentioned something about the drive itself doing it. How does that work? And can Scanner be used to trigger it?

 

UPDATE: ooh, I can edit my posts...nice. And, the Burst test works great (once I figured out I have to click Start (and Stop if I don't want the test to go on forever :) )

Link to comment
Share on other sites

  • 0

the "Ping" test creates a spike of activity, and doesn't keep the drives active, so wouldn't keep the LED active.  You need to use the "burst test" as this keeps the disk active until you manually stop the test (or restart the system).

 

 

As for "locking out" bad sectors, the term you're looking for is "remapping" or "reallocating".  And no, it doesn't.  In fact, the read pass that the surface scan does may actually trigger your drive to do this (in theory, it should).  
However, to force remapping requires writing to the sectors.  The problem with that, is that writing will eliminate the ability to recover that data.  Since the data integrity is the primary concern here, we don't offer this ability. 

 

IIRC, we do plan on adding this as a feature in the future, but again,... it is destructive. 

SpinRite and other utilities do write to the disk to force the remap. 

 

 

Also, generally, the disk should never come back with unreadable sectors. It should address these internally, and invisible.  So "bad sectors" during the surface scan indicate a "catastrophic failure", essentially. 

 

If you're seeing these, it would be in your best interest to stop using the disk and RMA it. 

 

 

 

 

UPDATE: ooh, I can edit my posts...nice. And, the Burst test works great (once I figured out I have to click Start (and Stop if I don't want the test to go on forever

:) )

lol, yup, that's what you want, and that's how it works.

 

It's great for stability testing.  It can reveal controller or cabling issues. As well as drive firmware issues (I've validated errors on a few SSDs using the burst test, actually) 

Link to comment
Share on other sites

  • 0

OK, this is an area I actually do know something about, probably because it's not really an area of tech.

 

For about 5 years, I managed a facility that manufactured circuit boards; not the finished boards with components, but the bare boards, ready to have components added. This industry really isn't techie. It's about machining and electroplating.

 

The raw material, fiber glass cloth covered with epoxy resin and copper sheeting on one or both sides, has to be drilled, punched, routed, and sheared to get it to the final shape. In the case of multi-layer boards, the epoxy-cloth layers are alternated with copper much like a layer cake. This allows for circuits through the middle of the board.

 

To get the circuits, the boards are masked in the pattern of the desired circuits, then electroplated with copper and solder. If the design calls for it, connection fingers are stripped of solder, then plated with nickel and gold. I'm sure you've all seen them on add-in cards for PCs. The extra copper sheeting is etched off the boards, leaving copper and solder circuits on top of bare epoxy resin.

 

Here's the real point: electroplating isn't an entirely exact 'science'. Variations in the flow of current throughout the electroplating solution in the tanks leads to variations in the deposition of material on the target (in my case the targets were circuit boards).

 

These days, all hard disk platters are electroplated. They, too, suffer from slight variations in the deposition of, in this case, magnetic material. These variations, over time, can lead to one area of a platter becoming non-responsive to magnetic impulses while the rest of the platter responds just fine. In other words, it's possible for one of more areas of a platter to become 'unreadable', or unable to take or hold a magnetic 'charge', while the vast majority of the platter remains fully functional.

 

If you get a few bad sectors, I don't think it's necessarily indicative of a failing drive. I think it's more indicative of variations in the plated material revealing themselves. That said, if a drive starts exhibiting many bad sectors, especially if there seems to be a cascade of them, then it could be a situation where nearly all the magnetic material on the platter surface has grown 'tired', and is no longer able to take and hold magnetic charges like it used to. Time to send that drive to platter heaven ;)

 

 

As for the remapping feature, I really look forward to that. I would like to humbly suggest that the user be given 4 choices:

  1. try to recover data from bad sectors;
  2. just remap bad sectors; don't try to recover;
  3. try to recover data from bad sectors, but only remap the unrecoverable ones;
  4. try to recover data from bad sectors, then remap both the recoverable and unrecoverable ones.

I would love to see these choices available in both manual and background modes. In other words, as a 'popup' window when interacting with the Scanner console? and as an option in the Scanner Settings (so they apply when the user isn't interacting with the console).

 

Oh, and thanks for the tips on Burst Mode. I will definitely give them a try, especially on the 2 Lian-Li enclosures.

Link to comment
Share on other sites

  • 0

thats some nice info! 

 

that said, normally, the drive's firmware should handle most of this stuff by itself.  Invisibly.  That's what the "reallocated sector count" SMART value is for :listing the number of sectors that have been reallocated due to various "issues".  

Additionally, IIRC, each drive is "calibrated" at the factory to account for the variations in the disk platters, reducing issues from that, as well.

 

Between these two things, you generally shouldn't see bad sectors, especially not in StableBit Scanner. And part of the point of the surface scan is not so much as detecting these issues, but forcing the disk's firmware to identify and fix issues before they become a problem. 

 

 

And again, to clarify, there is no software way to force reallocation.  What various software solutions do is read the data, and then attempt overwritting the sectors with information (usually/hopefully with the same info).  The reallocation occurs in the firmware.

 

Additionally, NTFS and other modern file systems feature the ability to do this for the file system, as well.  This is more of a stop gab, as the disk itself is where you want the problem fixed. 

 

The mapping and remapping of NTFS is handled by "CHKDSK" specifically.   And we actually call CHKDSK as the "file system scan".  Though, we don't use the /r or /b options, again, as these are "destructive" fixes (aka, you can't recover data from the affected clusters, afterwards).

 

 

The best way to fix ... well both of these is to write the disk.  A full format is a simple way to do this, as it writes zeroes to most of the disk. 

 

 

 

For the remapping, there isn't a lot we can do here. Writing to the disk is the ony real solution, and that covers the three options.  To "recover" or "remap" these sectors, we'd need to write to them. Period.  Otherwise, we'd need to call "chkdsk /b", which takes the disk offline for an extended period of time.

 

 

 

But the important part here, is that there isn't really a "great" solution here.  

And the only real option is to add the ability to write, as part of the "recovery" process.  And this is something that we have had on the back burner, for a while (we do plan on doing something like this).  

Link to comment
Share on other sites

  • 0

If I'm understanding correctly, and that's a big if, the way you would get bad sectors reallocated is to, repeatedly, try to write to the bad sector, eventually causing the drive's firmware to go, "oh, I see that's a bad sector.... here, I'll give you a good one to use instead". Is that about it?

Link to comment
Share on other sites

  • 0

So, it's been quite a day. I've read more techie stuff today than I can remember doing in a long time.

 

I was trying to find out what I could do about the 8 bad sectors that Scanner found in my main storage pool. Now, I need to mention that those 8 sectors are spread out over 4x2TB drives, so it's really only 2 sectors per drive (on average obviously; it is possible all 8 are on just 1 drive of the 4, but anyway...). I needed to find a way of locking out, or reallocating, or whatever, those 8 sectors, so I could continue to use the pool. So my hunt began....

 

In one forum I read about "badblocks". Apparently, this is a program that runs in Linux. The post also mentioned a Knoppix Live CD. I had to look up what that means. Basically, it sounds to me like it's an operating system that doesn't have to be installed; you can just carry it around and use it on different computers. So, I downloaded a version of Knoppix, after spending almost an hour trying to figure out which one to get. Man, do Linux people make it hard to understand anything: they make it all so complicated. Anyway, I was able to create a CD of Knoppix and boot up my father's old test computer with it. Then, I had to try to find out how to use badblocks. Well, that sure wasn't easy. Actually, I never did get it running, but more about that in a minute.

 

What I found out is that you have to know nitty gritty details about a hard drive before you can run badblocks on it. Turns out there's a desktop System Info app I can run to get basic info about drives, so I ran it. By the way, I noticed some really odd behaviours with the desktop. When you move a window, it distorts and wavers like it's being deformed. When you close a window, it breaks apart like a tiled wall being blown up. And there was other stuff. I couldn't help wondering if it was programmed by 12 year olds who thought gimmicks like these are somehow 'cool' (they're not). I don't know if these 'features' can be changed or turned off, but the fact that they seem to be the default actions says tons about their creators, and not in a good way. These guys probably think Rocky Horror is a current release movie :)

 

OK, so the System Info app still didn't give me the blocksize of the drive :angry: Apparently, I have to run yet another command in order to find that info. While searching that command (I honestly forget what it's called) I discovered that what badblocks really does is find bad sectors on a drive and make note of them in a file that the system can then use to avoid using those sectors. Sooooo, badblocks is of no use to me anyway, cause I won't be running Knoppix.

 

There's 3 hours I won't get back...

 

Back to the drawing board. Wow, this is so much fun :P  Courtney has mentioned chkdsk in a couple of posts. I hadn't seen it being run since the early 90s, when my father and I would use DOS computers (he set them up thank goodness) so we could play Duke Nukem. He would run chkdsk sometimes on those old machines, but he also said it was only minimally useful (that was my cue to erase any knowledge of it from my mind).

 

Turns out, the chkdsk he used to run, and the one that comes with Windows 10, are very different beasts. The new one, from what I understand of the stuff I read, is way more powerful, and effective, than that old one. No wonder Courtney mentioned it (thanks, by the way). Not only that, but it's enormously easier to understand and use than those Knoppix programs. So, after figuring out how to run a command window in administrative mode (a little obscure, but not nearly so much as Linux), I ran chkdsk with a few of the parameters I found in a Technet article about chkdsk (it was a great article; straightforward and pretty easy to understand, even for me). As I understand it, it will map out the bad sectors in the Windows file system, much like badblocks does in Linux, but it will be useful to me because it's in Windows.

 

I can also see now why Scanner doesn't (yet) have the ability to map out bad sectors: it appears that Windows itself is doing that work. Now I'm wondering if what Scanner should do is find bad sectors and then pass them over to chkdsk for resolution. I read about a "/spotfix" option for chkdsk that can 'fix' in seconds bad sectors that are in some kind of list or file. If Scanner could create, or update, that list or file...... Of course, it would only work for Windows 8 and 10 systems, but that should be most systems, right?

Link to comment
Share on other sites

  • 0

Badblocks is much like spinrite. Both will read and write to the disks.  It attempts to force the disk to fix or remap bad sectors on the disk. It doesn't communicate with the disk in any "special" way. 

 

As for the system info app, yeah, a lot of them are super showy.  "my epeen is bigger than your epeen" sort of thing. 

 

As for the block size of the drive, open up StableBit Scanner. RIght click or select a disk and click "disk details". This will tell you the sector size (and a bunch of other, really useful and really meaningless data :) ).

 

 

And yes, CHKDSK has gotten improvement over the years. But Windows 8 added some seriously kick ass new functionality (/scan, /spotfix).  It does a lot more, in a large part because NTFS itself does a lot more.  

 

 

And again, for StableBit Scanner, and these other programs (SpinRite, badblocks, etc), the only way they "mark" bad sectors is by forcing the file system (and disk) to write to these bad sections, and force the volume or disk to for it to repair or remap these sectors.  That's what it's doing. 

 

Additionally, "chkdsk /b" will force the file system to reevaluate the bad clusters on the drive, and attempt to repair them. This is why that flag can take even longer to complete.

 

 

 

But the take away here is: 

 

It still indicates an issue with the drive.  It may not be catastrophic, but it's still "bad".   But in my experience, this sort of issue usually gets worse.  

In some cases it can be "cleared up", but that's the exception, not the rule.

 

And the other part, is that if the disk is under warranty, the best option may simply be to RMA the drive and get a new one.  

Link to comment
Share on other sites

  • 0

I don't really think I would want to try to re-evaluate suspect sectors. I would rather just have them marked bad and move on.

 

There's no doubt that, over time, all drives will get worse. I don't need to know much about drives to say that. All I need to know is that all magnetic materials eventually lose their power. Anyone who's had a magnet laying around for several years should notice that it isn't quite as powerful after a few years as it was in the beginning.

 

When we're talking about electroplated disk platters, where the thickness of the magnetic material will be very thin, it seems pretty obvious, at least to me, that weak spots will develop over time. That said, I think a drive could still last for quite a few years after the first weak spots appear. I think drives would last even longer if the information on the platters is reinforced every so often. Nothing helps a magnet last longer than having its magnetism reinforced from time to time. :)

 

Can Scanner rewrite the sectors it reads? Is there a mode wherein it does that? If not, it might be a nice feature to add. In the meantime, if Scanner doesn't do rewrites, does anyone know of a Windows app that does?

Link to comment
Share on other sites

  • 0

/r looks for bad sectors and attempts to recover, /b re-evaluates, sorry. 

 

 

 

As for the weak spots, yup, definitely.  That's explicitly why the surface scan is good.  It runs a read pass over the entire disk.  Simply reading them should help.  And if the disk detects any issues, it should re-write the sections and "refresh" the bits.  

 

So while StableBit Scanner may not do this directly, it *is* triggering this behavior in the firmware.

 

(this process is actually referred to as "data scrubbing", IIRC)

 

 

As for Windows apps that do this? Format, diskpart's clean all. But these are destructive. 

I think that HD Sentinal *may* do this.  

 

There are other programs that do, but again, destructive. 

 

Also, SpinRite does this (but not a Windows app), as does badblocks. 

Link to comment
Share on other sites

  • 0

I get what you're saying, I think. However, maybe because of the drive, or maybe because it's in a pool, every time Scanner does a surface scan of this one drive, it finds these same bad sectors and lists the drive as Damaged. It doesn't seem to be triggering the drive to reallocate the sectors. One problem, now that I think of it (and which I just figured out), may be that the drive is actually 4x2TB drives configured as RAID 0 on a 3ware card. I didn't set it up this way; my father did. But, maybe, that's making it extra hard for the firmware to be triggered. I wonder if it's possible to have the drives connected to the 3ware card without having them in any kind of raid (I don't have enough SATA ports on the mother board otherwise). Maybe they would show up to Windows, and Scanner, as individual drives? Maybe?

 

Right now, I'm running chkdsk /r /scan /perf on the drive (it only took me nearly all day to figure out that these seem to be the best options to use, I hope). Anyhoo, it got through Stages 1 thru 3 with zero issues. It's now doing "Stage 4: Looking for bad clusters in user file data". So far, it's found 2 files it doesn't like, and they're now "queued for offline repair". I'm hoping that the repair will truly trigger the drive(s) to reallocate the bad sectors. Got my fingers crossed.

 

One of the files it found is a Vista update ISO so...... who cares :)  The other is a family photo from 2011. I'm not sure which one, but I should have backups, thanks to my father's VERY robust backup strategy (thanks Dad).

Link to comment
Share on other sites

  • 0

When that is the case (when it doesn't get "fixed"), that's because the disk is having problems reading from it, in general.  

 

This is when programs like SpinRite or badblocks would forcibly write to the sector, to try to force either/both the file system or disk('s firmware) to fix this sector. 

 

And yeah, the scan may fix these sectors.  But you need to run "chkdsk /spotfix" to complete the repairs (spotfix takes the disk offline temporarily, to fix queued issues). 

 

As for the Vista ISO.... I don't know, you might need that.  (I ... have a couple of Vista VMs running :P)

 

And glad to hear that he definitely did implement a good backup strategy! Seriously, you'd be amazed at how important that can be, especially when shit hits the proverbial fan. 

Link to comment
Share on other sites

  • 0

I agree. My father told me years ago that he, "has 4 to 5 copies of everything". He also said a number of people on the Homeservershow forums asked him to post his CMD files so they could implement something similar for themselves.

 

Right now I'm reading Terry Walsh's book, Building a Windows 10 Home Server. It covers something called File History, so I'm going to look at that.

 

What's a VM? Oh wait, I just looked it up... Virtual Machine, is that right? If so, it's kinda weird, running a computer inside a computer, like something ominous from science fiction.

 

Anyway, I just ran chkdsk /spotfix and it says it fixed the 2 problems it found on the drive(s). I'm running chkdsk /r /scan /perf on it again, to see if it finds any more.

Link to comment
Share on other sites

  • 0

Well, the running joke on Home Server Show, and We Got Served forums was "no such thing as too many backups".  Any time somebody mentioned backups, or drive failure or the like ....

 

Also, the 3-2-1 backup rule: 

  • Have at least three copies of your data.
  • Store the copies on two different media.
  • Keep one backup copy offsite.

 

As for File History Backup, yeah, that's a nice solution as well.  It's a versioned, file based backup.  But of just your data.  It works well, if you don't have a massive amount of data (like me).

 

 

As for a VM, yup, virtual machine. They're great for testing, tinkering, and the like.  We use them heavily, for testing. :)

Link to comment
Share on other sites

  • 0

Well, I ran chkdsk /r /scan /perf again, and this time it got to a Stage 5. I didn't see it do that last time. Anyway, apparently Stage 4 checks only sectors that are currently being used by files and Stage 5 checks the ones that are currently unused.

 

Stage 5 found 3 more areas (again, remember, this is over 4 drives, not just 1). It said it was flagging the areas for later repair. After it finally finished I ran chkdsk /spotfix. This took overnight to complete (well, I got tired of waiting for it to finish so I went to bed). It did complete though. I then went to Scanner and tried a Surface Scan. It sill said there are bad sectors and flagged the drive as Damaged. So, I went into the options for that drive and told it to Not do Surface Scans. At least, that way, I can still use the drive without Scanner preventing me from writing files. To be honest, despite what Scanner says, I haven't noticed any issues with the drive. And, since chkdsk has done its own scan of all the sectors, I'm hoping I'm fairly safe. And there's always the fact that I have quite a lot of backup. Between them all, I figure I'm not too bad off.

 

Oh, and I did something today that I think is pretty geeky: I set up OpenVPN on my (Dad's) Untangle box, something even he hadn't done. I had some missteps, and had to get help from the Untangle forums, but I was able connect up my Surface tablet to my phone's Internet Sharing and then use

Windows Explorer to get to the files on my new Windows 10 server. And, thanks to an article by a guy named pcdoc, I found out what Full

Tunnel means and got it to work. Now, when I'm away from home, I can surf the net as safely as if I was still at home. Pretty cool. Sorry, I know this is off topic: I hope that's OK.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Answer this question...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...