PureBasic Forums - English

I am trying to get the word out that there is a fatal flaw in the existing file structure for all drives an partitions. I can relate what is going on and what I see is happening, bur for various reasons, my posts and threads are then mostly hidden, removed, and I'm told to scale back the wording or be banned from that site. They ignore what I have to say because it does not fit in with the policy rules for that site.

For instance, I got banned from ubuntuforums.org because I did not wait for a question to be posted first before I acknowledged a problem existed, but went right to providing an answer. When I posted my own question then promply answered it, I got banned for teying to circumvent their forum rules. When I worked an alternate userid in to do both, they objected because my posts was too long and too hard to follow, and some people had a hard time following them so complained.

I just got banned by How0to Geek because I slammed Windows as a con job used to make Microsoft billions on profit and Bill Gates the richest man on earth. worth hundreds of billions of dollars. I said the game is over, because Miceosoft, despite itself, had to keep scaling up enough over the decades so that now it has to force people off its own prior products by terminating support, then selling you a replacement that gives you more features but actually reduces basic functionality in deeper and deeper menus.

It's blown it in five of seven releases, in that nobody wanted Me, Vista, Win8, or Win10. and only Win3.x, XP, and Win7 were deemed worthy. Win7 really is only good in respect to the lacks in Vista, Win8 and Win10. It's all in an effort to ensure they controlled the PC market and could keep milking the same cash cows -Windows from the OEMs and forced upgrades. and Office Suites that dominated the file side of the business. Since software does not wear out as metal and plastics do (Detroit's approach to ensure new car buys in the 50s and 60s that finally died when the foreign car makers finally got in and showed how cars should be built)

I just checked my own hanfle, and what I get on forums does get picked up by search wnginea. Even from years ago, because on forum. you fit in with the rest of the like-minded people interested in a topic or two. Not like a blog that nobody cares about and never visit. I even found reference I made to my free cloud account at pcloud.com where I try to make some of my writings public knowledge on forums that other would write objections to if I actually posted them directly.

I mran I make no secret of what I know or believe and am willing to share, but you have to care enough to seek me out, or be concerned enough about that topic. to ever find me or get my input. That's no way to fight a fire.

This is going to take a bit of programming mastery to resolve. It actually will require a standard's revision to make permanent. And it is about a single topic SUPERBLOCKS, which nobody has heard of.

The problem is, once created, superblocks, that are unique anchors in software on every drive, are never deleted. They change with every partition change, but if you move partition boundaries, particularly the forward boundary, the dozen or so superblocks written there at the leading edge offset remain behind. Being relatively positioned rather than absolutely position, they go adrift, and as you repartition over and over, more superblocks appear with each partition change as to its position among the others. They are never deleted, as the normal read, write, erase processes veer around them.

The only fix it to make them no longer appear to be superblocks, to overwrite the magic number they contain which makes them appear special. But if normal writes don't do it, then you have to revert to direct sector writes of replacement data, known as raw writes. Only how do you do this for real? I.m expert enough to know it can be done, but not expert enough to transfer that knowledge to code.

Now how do you know if I am giving you the straight facts or not? And how do you know that superblocks exist, or that they are doing what I say? Simple. Use a thrird party tool like testdisk and see the superblocks for yourself. Use the deeper search as find all the loose superblocks that mingle with those current. See how testdisk tells you when some block arrangements are bad, and there is only one that corresponds to your current arrangement and size structure, assuming your drive and partition structure has not disintegrated too far. And testdisk's findings should be supported by gparted at the GUI level, although gparted does not go into such detail. But neither tool, nor any other one of my knowledge, does anything to eliminate adrift superblocks. And that is wat needs being done to make drives truly reformattable. I've trashed drives I dare not ise now for fear of causing more content loss, due to overwrites, because at some point, the partition table is believed to be bad, or has been at least partially overwritten, by some process that went beyond its limits as defines by the standard.

This process is apparently programmed so that if it finds one or more supeblocks that appear connected and in order, it will revise the partition table for that one partition to match, but does not check the results for consistency with the rest of the drive.

I've had gparted and testdisk tell me that one drive suddenly had 1.3 TB of storage allocated on it, and that my swap partiton now occupied half the drive space of my third partition as well. Needless to day, partition 3 is totally trashed now as to its contents. structure.

That process programmer believed in the idea that if the number of agreements between this strring of superblocks (that were no longer valid), and changed the partition table accordingly, and that has deadly consequences.

Teskdist and gparted can identify the right lineup, but neither goes to the source of the dispute. They just return the partition to the state it was in initially, if that is still possible. it depends on the number and size of the new writes that followed. something you can't control due to logs and journals, even links and inodes being written to disk, continuing to take place, as well as virtual memory use if a swap partition exists on that drive

If you suspect bad files on other partitions, you may decide to delete then, but that just copies them to the Trash on your primary drive first, and that quickly pushes it to its capacity. And there is another flaw then exposed, in that a drive with 0 bytes free is rendered worthless until you boot up in a different way and clean out some of its contents.

You also find that there are certain folders and files that should never be copied from one partition to another mounted one. Some to exclude are /mnt, /media, /proc. /dev, /tmp. and likely /etc/fstab and /boot/grub/grub.cfg. You might want to exclude *[tT]rash* *.log *~ *.tmp and *.bak as well, as much of this is old news and no longer relevant. You want something to last, pick a different extension or location for it.

/media and /mnt folder copying are particularly dangerous (speaking for Linux of course, as I gave up Windows over a decade ago), because if you copy them to any mounted partition, or folder anywhere on your system, you probably create a loopback where you keep copying back more folders and files as they get added to the drive you are copying from. They don't just replace each other as the appear to have different paths between source and destination. To offer a simple example, if you have a /home folder and decide to create a /home/home folder as well. then you could create a loopback using a sudo cp -r /home /home/ command.

You will trash the target partition with overflow that goes on and on until every byte is full, as smaller and smaller files eventually squeeze through to fill the few bytes remaining. Since you are copying that partition or folder onto itself, but as an extension of itself, And it does not end until you run out of space. You can also create a loopback of you do a sudo cp -rf /home/home /home instruction, but it behaves differently.


I've had PureBasic for years, and I know how good it is and what it can do. But I have also learned the power of using bash and working with building blocks like find, grep, sort, awk, sed, cut, cat, less, trace, dir, ls, stat, teskdisk, gparted, df, du, lsblk, fdisk, and so on.

I have so much that is possible with these blocks of code that i can fit together quickly that the need for me to write compiler code just sort of evaporates away. It's all good code too, well documented, refined, and debugged, like a complete Lego set which which you can assemble most anything you want.

And rather than read up on each one and work out my own examples, I can quickly search out examples on the internet, so I just quickly move to getting it done without worrying about all the details. interpreter codes is thousands of times slower than compiler code, but PCs are tens of times faster than the mainframes I started with in the late 60's, and when you are that fast and we humans and internet that slow, who really cares for the speed differences?

But others do care and go the exte miles needed to learn all there is to learn so that they can carry matters a bit further. Many meet here, or know others that are of like mind. You are programmers mor than I m. and this may be your cup of tea to tackle and resolve with a simple tool that reads the testdisk log and take the necessary steps to do direct writes to specific ectors to wipe out the adrift superblocks.

gparted and testdisk together give you a simple way to make sure you have the right partition formation: Gparted lets you change partition labels, and testdisk appends these same lables as [label] when it writes the log. And testdisk gives you an all green flag when the partition table makes sense again, and its results then correspond to what gparted finds. gparted will correct tje partition table on a partition check, but the old superblocks remain. If you can write a bit of code to overwrite the adrift superblocks that are no longer valid, you have done what is needed for now.

But the real answer is in a different approach to partition validation. Superblocks should be absolutely positioned rather than relative. They should be on precise N intervals across the whole disk surface. Using x...x...x.Y.y...y...y...yZ..z...z...z... as an example, the ... represents som N interval, the small letters represent the superblock at the end of that interval, and the capital letters reflect the start of a new partition that can be anywhere inside any of those intervals.

Thus you know immediately where every superblock is by knowing N, and there is less need of a superblock magic number to say "I am a superblock. don't touch me". The fact that the superblock is precisely positioned with respect to other superblocks and agrees with the one before it, after it, or both together, makes it valid. The fact that its position is explicitly known in advance means it can be found and updated with no further assumptions. If it doesn't agree with either, look for what should be another superblock in that same partition and check it instead. You may be able to map out bad superblock sectors in the process, something that would be unknown otherwise.

I propose a new added standare of "a+" on front of existing partition types. Without the "a+", it is still the old structure. Convert it to "a+ext4" for instance, it is modified, partition, to use absolute addressing in place of the old "ext4" approach, where superblocks where located relative to when a new partition was created. If you skip unpartitioned areas, you need a way to flag these as unused or unallocated. You may want to simply copy the partition table here or checksum the partition table or take some other appropriate action related to integrity or security.

When I got into computers in the beginning, the big concerns were about data: Three criteria emerged: Integrity, durability, and completeness. The accuracy of the data was left to the individual. Now we have to worry about security as well, as well as distribution, availabilty, and who has what rights to do what with it. A failing at a previous employer is that anyone could change the data, there were no checks for consistency, and no validation processes applied. YWe had no real record of who was doing what, or if prescribed processes and procedures were being performed or had been completed, pr done by a responsible party and who that party was.

Maybe some of those concerns can be addressed now in a new standard, Some superblocks could home encryption keys for the partition that otherwise have no purpose. Or access codes embedded that you must know how to use to be allowed to read and write with to any part of the patition. Then we treat partitions in the manner of packets, where there is a head and tail that have precise meaning and size, but the contents are either what they are in agreement with the header and tail, but could be something quite different. You could devise your own partition format to your own sense of order and structure which you alone undersand, and which your hidden programs then manage as you see fit. To the rest of the world, it appears you are of a known type, vut you are nothing like they have encountered before.

The goal has always been for many is consistency avross the board, with standardized structures like folders, subfolders, files, and codes like EBCDIC, ANSII, Morse Code, international Code, unicode, written languages, math symbols, lefto to right (or right to left) then down for characters to be read, all for the intended purpose of sharing and communicating ideas and concepts.

The most significant advance in man's history was not fire as some suppose, but the eventual accumulation of many concepts brought together at last in the development of the reusable type printing press with sheets of paper and ink to use to make it work.

From that point on, the sharing of ideas and concepts moved towards the global stage, where many people from many places could now communicate en mass by printing multiple copies of their works and thoughts and discoveries. Now we are where we are in a few short centuries and no longer live isolated lives where the rest of the world was invisible to us.

That one invention became more, as electro-magnetic energy and its use moved us to the immediate transfer of information to distant points by wire, and broadcast wirelessly if more local.

We of course use the underlying concepts of sharing and dispersion as our goal, and use tricks ro hide or conceal what is there from unwanted eyes and ears. If you don't care to share, then there is no binding force on you to say you must be part of all this effort.

Create your own language or have no language at all, but then you mind will size up, because language is how we communicate even to ourselves. Language evolves from concepts, and language is how we relate back to those concepts. If I think 1+1=2, I know what I am thinking of, and I also remember that 1+1 does not equal 3. I know what 1 is, what + signifies, what 2 is, what equals means, and what 2 and 3 represent, and why the first is true and the second is false. and attempt to encase


When trying to explain matter to prople that are not familiar with the concepts and the language, you have to teach both at once. Just as I have tried to do here regarding syperblocks and their importance. If you read through what I said, you at least know where I am coming from, but you may not believe me if it is npt part of our shared experience. Id I saids I was green all over with golden eyes, that is not in your experience and you would not believe me. But if you know others like me or were one yourself, you would accept me at my word. Well, I make no such claim as I am neither green nor have gold eyes, and now you believe me because it fits what you know.

The problem with new words and concepts is gettting confirmation from somewhere beyond our experience. What is true and what is false? I offer two witness to what I claim: testisk and gparted, neither that I am associated with other than as a user. You can do as much, but you will put one of your drives at risk if you follow these instructions:

(1) Use gparted or other tool to partition, unpartition, repartition it into multiple partitions. Format each and maybe label it. each one different, and maybe copy some folders and files to it as well. Repeat these steps several times, and make sure the partitions do not match up precisely to what you had their previously, meaning different size partitions, different numbers, different formats, and different contents. When satiafied you have met this criteria, make use of that drive by installing and using multiple operating systems, one per partition. Make use of them. add more files and folders,

Move towards filling the partitions so that each one approaches capacity. Even do loopbacks to fill the partition to capacity as has happened to me when I was careless. Pefrorm updates and bing in more packgages, You will see. It is the repartitioning on other boundaries that serves as the tiping point. Testdisk will show scattered superblocks that don't go away, and it and gparted shoul still see a valid lineup of some or all partitons for awhile. but it won't last. They don't address and fix the problem, they just attempt to work around it.

Don't want to risk a valuable hard drive? Use an inexpensive thumb drive. Then you can share the experience and know the truth of what I say. A man that should know better claimed that in 15+ years of experience with creating and maintaining thousands of drives, he has never seen this as a problem.

But few people change partitions in number, size, and format as I do. This is a culmative problem that only surfaces if it means going beyond the morm, totally aviodable if you merely subdivide existing partitions onto more partitions, unless in the subdivision one of he smaller partitions get bled into by a portion of the string of superblocks from the parent partition.

Even then, there is a fair chance that the smaller partition created will not throw a fit if it is left more or less dormant. But later delete that small partition and inorporate it into something larger, and chances are, the old parent superblocks and its superblocks will now become a double whammy for you down the road as you make use of the new partition that replaced them both.

Then that same man of experience admitted he had seen a partition become corrupted when power was loss during an update. People don't realize that drives are serial devices, that it takes time to do each step completely. If a power loss happens, spmething may not complete in its entirety, or the signal may be distorted to the head, if say a lightning strike cause a power surge at that moment. You and I kno that software is always vulnerable, do we make copies and backup of just about everything.

But if we can't trust or depend on out drives and file structures, then what good are copies and backups? That makes them vulnerable too. What are you going to do instead, send everything onto the cloud or print it all out? What about their drives and the vulnerabilities of going online, or the drastic difference in transer speeds? How many hours would it take to back up a terrabyte of data to a cloud server, pay the going rate (at your expense), then being it back on demand? A cheaper course is to get another 1TB HDD for less than $60 off Amazon.com, but then you are back to repeating the same problem that sent you there in the first place.

If you go with the new 1TB HDD drive approavh, just reformat the existing partiton that is on it. If you want a swap partition as well, shrink the original partition down from the tail end using gparted, then add the new swap partition in the space dleft unallocated. You want another partition there as well? shrink the original partition down enough for both as well. What you don't want to do is move any partition at the front end or not replace it with a partition that does not align at the front end with what was there before. This should be enough to keep you safe from this particular problem. It is not a fix, but it is a bandaid.

































Two complications immediately become apparent: Relative addressing and assumption thatr the magic number identifies a valid superblock under all circumstances, and Corruption returns because only the partiton table is repaired, and other changes expand further. I've been over 4 months running this problem down, and the only tool to tell me enough was testdisk. It's log is the basis for which sectors need change. So knowing the sectors, what is the method of getting some changed?

As long as you only chage the format and leave the .

Sorry,

to many to read no code to fix -> no interrest.

Where is the relevant PB "Coding question"?

Prolix!

no clue what you want to say here. if you want to make your own filesystem with fully relative structures and adressing without relying on magic numbers or absolute adressing - then simply do it, instead of trying to make others do it with their filesystem. but what would be the point?

i have written commercial tools which process filesystems and partitions for many years, studied computer science in an operating systems lab and still dont see the problem you have with "superblocks".

you also dont need any "testdisk" to reveal them, you can simply use bios calls or write your own ata driver code or just lock and open the harddisk device to access each physical area (limited to what the controller allows you to see

Then that same man of experience admitted he had seen a partition become corrupted when power was loss during an update. People don't realize that drives are serial devices, that it takes time to do each step completely. If a power loss happens, spmething may not complete in its entirety, or the signal may be distorted to the head, if say a lightning strike cause a power surge at that moment. You and I kno that software is always vulnerable, do we make copies and backup of just about everything.

a decent filesystem fixes that by enforcing the order of data it writes. It can get slightly complex to achieve, but the worst thing is that it is usually physically slower.

See for example https://en.wikipedia.org/wiki/Transactional_NTFS or look for transactional filesystems like Reliance (http://www.datalight.com/products/embed ... ance-nitro)

ps: the purebasic forum is probably not the place for this

pps: usually you simply use one of these, to fix the whole problem: https://en.wikipedia.org/wiki/Uninterru ... wer_supply

Two points I should have made: (1) I only used several distros of Ubuntu when this began happening on my maclines. It first showed up under 14.04 LTS. It was moe evident under 15.10 and 1604. But it was totally out og control under Uguntu-Gnome 16.04.

Now these were progressive stages, and the problem may have had nothing to do with the LiveCDs I was woring with. I did not test other distros as I was only attempting data recovery, a still incomplete process. It is only with the help of testdisk that I had enough visibility to make the superblock connection. c That only occurred last week.

Not sure about Ubuntu, T downloaded the SparkLinux Rescue CD. as it was Linux-based. Most recovery disks seem to be Windows-based, and I would expect Linux to be more Linux-focused, and Windows more about FAT16, FAT32, and NTFS. I could be wrong, but there is a gap when it comes to specific details about how thorough theses packages are.

On the second point, brought up in a reply, I thought it would be apparent that I wanted to know how PureBasic could be used to perform direct writes to sectors. It is based on C, and I found mention that C has direct write capability to the drives.

But it was alo to run up a warning flag, that if you have repartitioned ypor drives, you may be vulnerable. This problem is not specific to Linux, as superblocks are used as part of the total drive structure, common to all partition types.

Hoping to avoid too much aggrevation of the situation, I tried several partition types. There are a number of gaps there between what gparted supports and what the installer allows. I found that gparted was better at formatting partitions than the ibstaller is, and ext4 blows it with ext3. But ext3 is the only one that takes long enough to give the impression of a long format. Whether this took care of overwriting superblocks in the partition is unknown, as ext4 proceeded to make hash of that partition and more. That drive is very unstable now, and my last data store that I need to recover.

So my current problems could be related to Ubuntu, Debian, or Linux in general.
Or it could be more generic. in that it could show up under Windows.

Could orphaned superblocks be an unexpected consequence of switching from full format to quick format? I kind of doubt it, but thenI don't know at what point new superblocks are created, or how extensive a full format would be. It might ignore orphaned superblocks within the partition.

oldefoxx wrote:I am trying to get the word out that there is a fatal flaw in the existing file structure for all drives an partitions

what i find really strange is if there is such a "fatal flaw" for "all drives and partitions" why are you seemingly the only person talking about it? why isn't it all over the news at arstechnica or whatever, and why are you getting banned at so many sites as you say for simply talking about these "superblocks" (these what? is this your invention or an established thing?) if it was an issue of genuine concern? something doesn't add up!

Dealing with filesystems and their structure, can get complicated. The ability to access raw blocks of a hard disk, in an attempt to rescue data from some kind of logical data corruption (no matter if it was the partitioning data or a filesystem structure has been affected), is a tiny step.

What you need is some kind of software like http://www.easeus.com/data-recovery-software/ or the service of commercial data recovery companies. As far as i know, the term "superblock" is commonly used in the unix world and you usually dont see it being used in windows fat12/16/32/32-ext filesystems. The concept of superblocks, exists in most filesystems beside some scientific projects like amöba where disk storage can be very different.

Forget about doing data recovery yourself, it would take you way too long to get there in time.

In theory you can write that software in pretty much any language. I wrote a lot data recovery software in pascal years ago and some tools even in purebasic. Look for int13 or "extended int13" with the search function of the forum to find stuff like http://www.purebasic.fr/english/viewtop ... 35&t=53834

http://www.codeproject.com/Articles/283 ... sk-Sectors

I will look for some old examples in purebasic to get it done, but i havent seen those for years - so no promises

Thanks people. No answers yet I see, but at least it is getting some attention. I've stayed away a few days to let this simmer a bit, and get a broader response possibly.

As to dome og the points raised: Testdisk reports both FAT32 and NTFS superblocks as well. So they go deeper than you realize. Potentially, there is risk to everyone here it seems.

The risk can possibly be due to one faulty process. Finding that one process would be a nightware. Maybe bringing the problem out in the open would cause someone in the know to takes a harder look at their code, or realize they missed something in what they wrote.

The fact that it isn't reported or not known to be a problem can be attributed to many things:

(1) If you don't repartition your drives, or you just lay new partitions down in the same locations as old ones, you should have no new problems.

(2) If you have a drive or partition corrupt on you, you may do a data rescue and just go on from there. If it repeats, you may assume you have a bad drive and replace it.

(3) If you split a partition into more partitions, an option with LiveCD, as long as the original partition starts at the same place, then new superblocks should replace the old ones.

(4) If nobody else says there is a problem, then there must not be a problem, right? I just went through an episode with my ISP because a friend is handicapped loses her handicapped status with them month after month, forcing her to refile for her handicapped status and a manual adjustment to here bill. Over and over this happens.

You think she is the only handicapped person effected? Obviously not, but she doesn't think on those terms. Neither do the individuals in Customer Service that have to help her every time she calls. Nobody has ever tried to open a big report because that is not their job, they did not think to do so, or they don't know how and won't bother to ask.

The ISP does not even have a point of contact on their web page other than sales, and no chat session either. Used to be that you could send an email on the blind to support@<domain> or webmaster@<domain>, but that was just a sort of understanding of sorts.

If I new how to write the code I would write the code then share it. Obviously I am not that good. And as one responder commented, it would probably take me too long to learn how. Right now my wife is in a recovery center where she will be until her fractures heal enough for her to come home. Extimates are that this make take two months, as her bones are fragile. She's 80 years old, and has many of the effects of aging put on her. She wants me to ready a laptop for her that she can use there, which is what I am doing.

She wants Windows 7, and I agreed. So I stripped the hard drive and did a fresh Windows 7 Ultimate install. Things wen well for a couple of hours, but a self-check turned up 15 errors. I ran the fix, but as soon as it was fixed I had 24 new errors. These it would not fix unless I ordered the Security Center.

The price was too high, and I tried to leave the page twice, and each time the price came down. I finally went for the price offered, and that brought me into contact with Microsoft support. The contact got me registered with a download, which gave him remote access to the laptop. I watched him work as he explained things to me, and what he told me was not good.

I had over 1400 additional files added to the laptop in that rwo hours, and when he checked my security features, they were all stopped r removed. He could even see someone on my PC remotely as we were sitting there, coming through my modem and router. And that is with the router rigged for encryption and password access.

It was so bad he said he was going to have to turn it over to a special team to clean up, as he had no idea how far it went. But for him to do that, I was going to have to pay a one time charge of $247, which would buy me support for a year. My wife had paid someone that same kind of money to clean her desktop some months back, so now I knew who she was dealing with.

In her case, they then wanted hundreds more to add malware protection, but this she refused to do, and they kept at her by phone for two days until they got the message. I hung up on them, ignored their calls, and decided my wife was going to have to learn to like Linux. We were not putting out that kind of money just to have a computer she used for email to access accounts from, and to play a few computer games. I told her by phone what had happened, and she was in pain and on pain killers, but she understood enough to agree. But whether she will later let me redo her desktop could be another story. We will see.

Meanwhile I have to decide on a distro to move to. As my bad efforts were all Ubuntu-based, I decided to try an alternative. SparkyLinix Rescue locks up, so I would skip SparkyLinux. I wabted it to be gnome and debian, but not ubuntu. Ao I decided to look a OpenSUSE, which others seem to like. And it is big, 4.3 GB in size, and will take hours to download. But there is an advantage to that. Unlike Ubuntu, it should not require you to be online to complete an install, somethig I never liked about ubuntu. So we sill see what we will see.

I'm done with Windows, unless I can install and run it safely as a client under Oracle's Virtualbox. Something I've done in the past, and might still be possible. If not, I should be able to get Embird to run under wine, which is the one thing my wife won't give up. Others say they've done it, so I should be able to as well. I might wvwn fwr hwe favorite mahjongg game to run under wine. They she should have no quarrel with getting her desktop off Windows for good.

There are Windows lovers out there, and if that is your thng, I;m not going to try and talk you out of it. It is, after all, your PC and your life. But what are you going to do when Microsoft falls? It will, you know.

There is no high end PC market to speak of any more - everybody is going small as laptops and desktops are overkill for their needs. The small end market is dominated by Android and web apps, all free. Intel is feeling the shrinking market, and look at what extremes Microsoft is going to make a dollar or two, like pestering you for days to close a deal with them. They aren't fixing the weaknesses in Microsoft, they are making money off them with all these added charges. If you don't pay up, you are going to get hacked. I got hacked in under two hours, something that has never happened before as far as I can tell.

The bad thing about fixing problems after the fact is that you aren't kept from being attacked to start with.

Look at the numbers: Windows is the biggest and least protected target in software out there. You have it overdoccumented to the point where its vulnerabilities can be seen and shared. You have a dozen releases to target, most alike, whereas the distros and flavors of Linux runs into the hundreds. People that learned Windows think there is no other way, and others think it is secure because they don't see the insides. With Linux, you get daily or weekly updates and upgrades, and you decide what you want. With Windows, upgrades have to wait for the next relase so they can seel that to you, nd updates are mostly bug fixes in Service Packs. I'm sorry, I just don;t see the logic to staying with Windows, especially as it can amount to big money, it has essentially failed with 4 of seven releases now, and nobody likes their sence of direction in terms of functionality and utility. The failures were Me, Vista, 8, and 10. And look how vulnerable 7 has turned out to be.

That makes XP its last good product. But XP was just 2K with a lot of added licensing and restrictions imposed to ensure you were limited to deployment, and getting any updates if you were not properly registered. I worked 2K for years because it was all right, and there were no restrictions on its reuse.

Now these are personal opinions of course, but I've been around long enough to have learned a lot first hand. So I consider them informed opinions, but you may not agree.

I will keep watching this thread, hoping for something specific as to getting rid of unwanted orphanned superblocks. I will look at the links offered, but so far, nothing gets to the superblocks themselves. Testdisk still finds them where they were before. As I have corrupted drives now galore, it should be easy enough to check to see anything works. It just takes time though, and my wife's needs come first. I'm just using the OpenSUSE download time for this and other things. I'm gone again.

Rumination, ad nauseam!

tl;dr

c ya,
nco2k

I often start reading these posts but after one paragraph or so I loose my interest. The text is far too long. I need years to read it.

oldefoxx wrote:No answers yet I see, but at least it is getting some attention. I've stayed away a few days to let this simmer a bit, and get a broader response possibly.

As to dome og the points raised: Testdisk reports both FAT32 and NTFS superblocks as well. So they go deeper than you realize. Potentially, there is risk to everyone here it seems.

Lets assume that you keep talking about this software here http://www.cgsecurity.org/wiki/TestDisk.

This kind of software, usually scans for known filesystem structures and it fully depends on the software how good the results are. It could give you tons of faulty detected structures or "superblocks" - it fully depends on the author(s) of the software and i doubt he is reading here.

oldefoxx wrote:I will keep watching this thread, hoping for something specific as to getting rid of unwanted orphanned superblocks. I will look at the links offered, but so far, nothing gets to the superblocks themselves. Testdisk still finds them where they were before. As I have corrupted drives now galore, it should be easy enough to check to see anything works. It just takes time though, and my wife's needs come first. I'm just using the OpenSUSE download time for this and other things. I'm gone again.

A Hard disk stores data in a huge sequential stream of data which are logically seperated into blocks. In the past, those "blocks" where equal to a single sector, but as hard disks get larger, one way of handling larger sizes, is to make those single units larger than one sector, so the disk adressing still works with 32- or 64-bit numbers. The hard disk (or more specifically - the controller which does the communication with your hardware) gets a number to read a sector and then uses that number, to position the physical heads above the area, spins the disk and reads the magnetic stream below the head, to decode it and deliver the sector to the software which requested it.

For simplicity, lets say that every sector is the same to the physical hard disk.

The first software to access your physical hard drive, is usually the BIOS as it starts and tries to "boot" your PC. The BIOS reads the first sector of your first hard disk and treats it as a MBR (master boot record). That's the structure which contains a tiny area of code, the partition table and some bytes for a signature. At that point, the information of the MBR will give the logical structure of the sequential stream of sectors on the physical hard drive. So the BIOS sees where partitions start and what type of partition they are.

Filesystems are stored inside those partitions. The partition ID from the MBR often implicitly gives clues about the structure of blocks inside that partition, because most partitions are pretty simple and just contain a single specific filesystem. The BIOS proceeds to load the first block of the "active" partition and exeutes it (the first sector of any partition is the BOOT sector and one of the most important structures on a partition).

What i try to say here, is that anything else like making a single sector something special like a "superblock", is just done by some software and technical convention like a RFC or a . Physically all sectors are the same. Some filesystems have Node blocks and "supernode" blocks and most apply all kind of logical structure to its content (directory blocks, information data about the inner structure of the partition like MFT, swapspace, FAT, hibernation data and others). Thats likely what you refer to as a "superblock". They might be crucial to a filesystem, but to the physical hard drive and its controller, they are just the same like any other block.

If you format a partition, you can do a quickformat. That quickformat will just write the new filesystem information and not overwrite the existing sectors with "fill bytes". If you quickformat a partition and use a tool like testdisk, you might find structures for two partitions on your hard disk (usually if the old partition and the new one that was quickformatted had a different size or location). Quickformatting is convenient and we likely all do it, but for later recovery processes, it can confuse any recovery software (like testdisk).

Dont try to use recovery software like testdisk, unless you have to recover something.

If you want a more secure data storage system, look for RAID and use a UPS plus use a good Backup strategy.

That will reduce the risk of losing data by a lot, but even 99.9% means that a loss could still happen, its just very very unlikely to happen again during your lifetime. Remember that all of these will just combat physical problems, a logical problem like a trojan could still infect your computer and wipe out all data. A really clever one might infect you and sit dormant long enough to be in all your backups too. Unlikely to happen, but not impossible.

But the most important thing - all the best for you and your wifes health

that last post actually was the first superblock in that thread that i could read to end.

nco2k wrote:tl;dr

c ya,
nco2k

Neither Ubuntu nor OpenSUSE recognize tl or dr. Looks like you have something going under Windows. Some more details whould be greatly appreciated. This muxt be custom code you are referring to, not generic to Linux. If it does what you suggest it does, that would be really great.