jaclaz

Yes and no. It is important to understand the difference between a boot manager and a boot loader, and since most of the tools are at the same time BOTH, it is easy to get confused. A boot manager is something that can chainload (or pass control to) a boot loader. A boot loader is something that directly loads an Operating System. The default boot manager for NT through Server 2003 is NTLDR. NTLDR is also the one and ONLY boot loader for NT through Server 2003. The default boot manager for Vista and later is BOOTMGR. BOOTMGR is also the one and ONLY boot loader for Vista and later. NTLDR is at the same time a boot manager and a boot loader FOR NT/2K/XP/2003 ONLY. BOOTMGR is at the same time a boot manager and a boot loader for Vista , 7 and later BUT among it's boot manager features it has the capability to chainload the NTLDR, using it as a boot loader. See here (that includes a nice set of graphics): http://www.multibooters.co.uk/multiboot.html Simplified, right now to boot XP you use only NTLDR, and since NTLDR by itself cannot chainload BOOTMGR, we use grub4dos as a "man in the middle" (as an intermediate boot manager). If you change the primary boot manager from NTLDR to BOOTMGR you can get rid of grub4dos, since BOOTMGR can chainload NTLDR "directly". In order to do this you will need to have a copy of the BOOTMGR and of the \boot\BCD on your first disk and change the PBR on the Active partition of it to load BOOTMGR instead of NTLDR, then add to the \boot\BCD a choice to load the NTLDR (if - by any chance - you use Bitlocker it will be needed to change also the MBR code). The MS tools to do the above are bootsect.exe and bcdedit.exe, as they are a bit complex to use, ask before attempting to use them and if needed I will provide you with the exact procedure . jaclaz

Well, you cannot use NTLDR to boot Windows 7, which is what you are attempting to do with this line: Now, be nice . Start again. You want this BOOT.INI: and add to the first disk (together with boot.ini), this menu.lst: and of course a grldr file, you can get it from the latest 0.4.5c version: http://grub4dos.chenall.net/downloads/grub4dos-0.4.5c-2015-04-08/ http://dl.grub4dos.chenall.net/grub4dos-0.4.5c-2015-04-08.7z Theory of operation: Bios boots the MBR of first disk. The MBR code loads the PBR of the Active partition on first disk. The PBR code loads the NTLDR (residing on the active partition on first disk) The NTLDR loads the BOOT.INI choices. You choose "Grub4dos to Windows 7". The NTLDR chainloads the grldr file (grub4dos). The grldr loads the menu.lst choices. You choose the "Grldr Loading BOOTMGR" choice. The grldr looks in all volumes on all disks if he can find in the root a file "BOOTMGR". Then it chainloads the found BOOTMGR. The other entries (that should all work) are other variations with the same final goal of chainloading BOOTMGR on the first partition of second disk. The BOOTMGR loads the choices in \boot\BCD (possibly just one in your case) You choose to boot to Windows 7 (or wait until the timeout in \boot\BCD) and BOOTMGR chainloads WINLOAD.EXE (which is what actually boots the Windows 7). If you prefer, when you boot and choose the "Windows XP", you are using NTLDR as both primary boot manager and bootloader, when you boot and choose "Grub4dos to Windows 7" NTLDR is used only as primary boot manager, grub4dos becomes the secondary boot manager and BOOTMGR becomes tertiary boot manager and boot loader. I know it is confusing , but you should have no issues in following the above with some patience. jaclaz

Sure, and to change boot one would need to change drive order in BIOS, which brings us back where it all began, i.e. here: http://www.msfn.org/board/topic/156859-drive-order/?p=1005033 which, starting from here: http://www.msfn.org/board/topic/156859-drive-order/?p=1005435 details the grub4dos usage to have each OS "independent". jaclaz

Wait a minute. Try EXACTLY the example given by Joakim http://reboot.pro/files/file/237-runassystem-and-runfromtoken/ jaclaz

Try insulating the "other" set of contacts or try with the PCB completely detached from the drive, see here: http://www.msfn.org/board/topic/128807-the-solution-for-seagate-720011-hdds/?p=1079049 jaclaz

Appalling. Thanks for that link, really interesting. All in all I would then see even less reasons to run an ugly, bloated OS as a host for the virtualization, at the end of the day, I would rather have (say) QNX running a VM , hey, wait, seemingly it is exactly what the good guys at QNX are after : http://berryflow.com/2015/02/blackberrys-qnx-launches-hypervisor-1-0-vmm/ jaclaz

Well, if you install first XP, the volume on which you install it (let's say the first primary partition of first hard disk) will get drive letter C:. When you install the Windows 7 that volume will remain C: and the volume (let's say first primary partition of second disk) will get a "later" drive letter. The booting until you have only the XP installed will go: BIOS->First Disk MBR->Active Primary partition PBR (this will be C:\)->NTLDR->BOOT.INI->Windows XP once you will have installed the 7 it will go either: BIOS->First Disk MBR->Active Primary partition PBR (this will be C:\)->BOOTMGR->\boot\BCD->XP choice->NTLDR->BOOT.INI->Windows XP or: BIOS->First Disk MBR->Active Primary partition PBR (this will be C:\)->BOOTMGR->\boot\BCD->7 choice->Winload.exe (on second disk partition, possibly D:\)->Windows 7 If you have only one primary partition on the first disk that will be C: under both OS, whilst the first primary partition on second disk will be D:, again on both OS. As always, using a few tricks and/or a third party bootmanager this can be changed if this is not the expected result. jaclaz

Yoiu convinced me . I am pretty sure you have enough grounds to sue Jumper . But seemingly Drugwash only tried to make you aware of current status of the thingy, answering your question, maybe he deserved not the lecture on GPL compliance and opportunity or using version control. jaclaz

Yep , and I would go further. Tablets (and smartphones) are contributing to killing syntax and grammar of languages or force the user to write in an invented, simplified, language, not suitable for any meaningful purpose. This effect goes almost unnoticed in English because it is an alphabetically simple language, but I can assure you that in languages like Italian and French (and I believe German also) where some "special" characters are needed the result is very noticeable. Till the simplified language was confined (understandably) to kids and to SMS's it was of course fine, but you simply cannot write properly (or it takes you such an effort that it is highly counterproductive) on the "simplified" screen keyboard of a smartphone/tablet. Little by little everyone starts removing accents, proper punctuation, etc. and the result is appalling. About this: as I see it the issue is not with tablets in themselves (which I believe when used for the right scopes are nice little thingies), but in the (false) perception (of course fueled by the advertising and what not) that they are a replacement for a "real" computer, while in reality they may represent an extension to it, or rather a different commodity. I am typing this on a very normal desktop, I have before me a 22" display which is set at the right distance to allow me to see the text I type comfortably, I have under my hands a "real" keyboard (IBM M ), I don't need to look at it to type (and it doesn't cover half the screen area), if I notice an error I have arrow keys that allow me to move the cursor back where I need to make a correction, I have both a "backspace" and a "del" key, I can use SHIFT+ arrow keys to select text, etc., etc. I have a teeny/tiny laptop, which I use when I have to do some work on the move, which is in my perverted mind an extension to the above main PC, and while less convenient for typing, and forcing me to a not fully comfortable posture, is OK for the occasional activity. My mom has an iPad which is an exceptionally good device for what she uses it for (e-mails, some Facebook, a few games, listening to music, looking for info, timetables and the like, searching translations or the occasional definition on dictionaries, researching this or that topic on Wikipedia or similar) which I happen to use from time to time and as long as I do those same activities on it (excluded Facebook) it is a pretty nifty little thingie, but I would not even dream of using it to work with it on (say) an Excel spreadsheet. Now: Would I possibly buy "a" tablet? Possibly yes , my smartphone (which is anyway a very "large" model) has a far too small screen to allow me to read comfortably, and even if my netbook is a very small one it is anyway not "perfect" in portability, I would find an use for the tablet to have it (say) near my couch at home ready if I want to check some info on the internet or have it with me and use it as an e-book reader in a waiting room. Would I possibly buy specifically a "Windows" tablet? Not really , simply because there is no way I could use it as a replacement to any of the other two devices for the use I have for them, and if all I would use it for is some internet browsing, occasional e-mail checking, etc. there is no real reason why I would prefer having a Windows OS, both iOS and Android do that nicely and they are both all in all "user friendly enough", and I am pretty sure that even the stupid Windows 8/8.1 (and the next 10) would do as well nicely for these trivial activities, if I had to choose (being notoriously cheap) I would probably go for the less expensive thingie that would fulfill my requirements which may (or may not) be a Windows based one. In a world where everything will be (before or later) internet based the focus will probably shift from OS to browser, and just like it is now with Chrome vs. Firefox vs. Safari vs. IE vs. Opera[1], etc. the personal preferences would be more aimed to having available your browser of choice on the given hardware platform and the OS will probably fade away in the background. jaclaz [1]For the record after having fought uselessly against Chrome, I have Opera (mini) on my Android smartphone and it is not too shabby.

@Formfiller Sorry to say so, but you are IMHO missing the whole point. It was noticed alright, as it is talked about in the actual release blog post for that build, here: http://blogs.windows.com/bloggingwindows/2015/04/22/windows-10-technical-preview-build-10061-now-available/ The issue here is that a build with such an ENORMOUS issue should have never have come even NEAR a release, and that IF that happened the release should have been canceled, instead of "downsizing" the issue as a mere bug (even if described as "a bit painful") and providing a (lousy) workaround: Imagine Toyota (I am taking Toyota as an example as they have built for themselves over the years a fame for their reliability and pursuit of excellence) engineers presenting a car at specialized press for a drive test (i.e. well before going into mass production and selling the car, just presenting it as a "preview") telling the journalists: "There is a bit painful bug in this car as it doesn't have seats, as a workaround you can use these fruit crates". Unconceivable. jaclaz

Internal testers? WHAT internal testers? http://www.imdb.com/title/tt0072431/quotes?item=qt0484668 More seriously, I don't think it is intentional, I believe that the level of internal testing (if any) is so low that such a big tree lobster could go unnoticed, most probably the "internal testing" (again if any at all) is surely focused to LOLz and Facebook and Bing and not to actually "working". I mean, if - say - a lesser known program like (in the US) TurboTax doesn't start, it is a bug, if their own Office suite programs don't it means the NO productivity program has been tested. But a mistake can happen to everyone , I have nothing against the good MS guys that were involved in the (completely failed) testing, what I am saying that Gabe Aul calling this "a bug" is totally inappropriate , any senceful "project manager" would have HALTED the release instead. Imagine - say - Ford selling the Pinto knowing that it was dangerous in case of accidents .... NO, wait, I got the wrong example : http://en.wikipedia.org/wiki/Ford_Pinto#Allegations_and_lawsuits but anyway that happened some 40 years ago, maybe something has changed for the better in the meantime... jaclaz

Sure , I was just highlighting how things should be called with their name, IF that is a bug, it is a "tree lobster": http://en.wikipedia.org/wiki/Dryococelus_australis jaclaz

The normal XP booting: BIOS->MBR->PBR->NTLDR->Choices in BOOT.INI: 1. arcpath to Windows installation 2. PBR/bootsector copy pointing to the CMDCONS (Recovery Console) The way the Recovery Console is loaded is detailed here (JFYI): http://reboot.pro/topic/2362-makebscmd-alpha-release/#entry18878 The normal Vista booting: BIOS->MBR->PBR->BOOTMGR->Choices in \boot\BCD: 1. path to Windows installation WINLOAD.EXE The Vista booting installed after the Windows XP: BIOS->MBR->PBR->BOOTMGR->Choices in \boot\BCD: 1. path to Windows installation WINLOAD.EXE + ANY ENTRY in BOOT.INI which is NOT an arcpath i.e. 2. PBR/bootsector copy pointing to the CMDCONS (Recovery Console), if needed pre-modified with the "cmdcons" string as per given reference. The real issue you may have is if you use the same volume to install the XP and the Vista (because of possible conflicts with paths), if you install (say) the XP to C:\ (first primary partition, active) and the Vista to (still say) D:\ (either second primary partition or logical volume inside extended) you will be fine. In any case, you can still add to the BOOT.INI (to be parsed by the BOOTMGR) a copy of a PBR loading the "old" NTLDR or add to it grldr (grub4dos) and from it boot *anything*. jaclaz

I don't get it. The original XP Ramdisk.sys[1] does work nicely though it needs a registry "patch" and it is limited to 512 Mbyte (not because of itself, but rather because of some limitations in the loader(s)) the size issue when booting is seemingly with the version of the loaders, see: http://www.911cd.net/forums/index.php?showtopic=25389 http://reboot.pro/topic/9474-busting-the-myth-about-ramdisksys-xp2003/ As often happens, it is not clear what is your "final GOAL" (NOT the way you believe you should use to get to it). Care to detail that (the final GOAL)? It is possible that it can be reached through other paths. jaclaz [1] Not really, I am actually lying SP0 and SP1 were never AFAIK/AFAICR tested

Well, no . That is NOT a "bug". A bug is something that by definition is small and thus not immediately evident, that is an essential, basic, fundamental, requisite of a graphical OS, being capable to start a program from the UI part designed to start programs, heck, it is called "Start menu", that is a LARGE (PINK) ELEPHANT! not something that would go unnoticed... jaclaz

Yep, a VM would be a perfect test bed. As a side note 99.9999% of (please read as *all* ) "recovery partitions" are Primary volumes on MBR disks (and of course all volumes in GPT are primary partitions) so there won't be issues in the suggested "make a copy of the disk from sector 0 (the MBR) up to the beginning of first partition/volume", but in the extremely rare (please read as non-existing) case of a recovery partition being actually a logical volume inside extended one would need to additionally save the whole chain of EMBR's. jaclaz

Hmmm, never heard about that. Well, you never know. As a matter of fact the WIMBOOT (or WOF driver) is something that has been added in Windows 8.1, basically it is a "simple" enough approach, it is a kernel driver that can interpret a .wim file as if it was already an applied volume, so it does not in itself constitute anything particularly "complex" to deal with. But we are still in the earlier misunderstanding. Microsoft advocates the use of a "recovery environment" which is based on a "normal" .wim, the WinRE.wim, starting from Vista, the situation AFAIK is/was the following: with Vista the method was recommended but very few OEM used itwith 7 it was strongly recommended and a few OEM's ditched their own "custom" recovery environment and adopted the MS waystarting with 8 the "recovery environment" became mandatorystill they are two different things, one is the "recovery environment" (which is a kind of "built-in PE") and one is the "recovery partition" or the "recovery media" that you can create on DVD(s) or on a USB stick. What is now compulsory is the presence of the WinRE, the OEM is free to use the Recovery Environment to apply/restore the mass storage device to "factory state" but remains free to use a completely different, custom, approach to obtain the same. A number of low-cost/small SSD sized tablets may use Wimboot lately as the advantage it has is that the senseless amount of bloat a Windows install represents is strongly reduced due to the WIM format compression, though this only works initially as each and every subsequent update is going to be stored "flat" on the device (and it seems like the functionality will be removed in Windows 10), see: http://www.msfn.org/board/topic/173786-whats-up-with-the-bevy-of-optional-windows-81-updates/ From the contents of the USB drive you have, it does seem like a "normal" install/recovery, probably aimed to be saved on both DVD's and USB stick, if you check the size of the .swm's they should be compatible with DVD sizes, the presence of the reagent.xml file is "standard MS", while $PBR_DiskPart and $PBR_ResetConfig are "non-standard" AFAIK. So it is possible that the USB stick you have is a "mixed mode" making use partially of the MS approach and partially on some OEM customizations, the presence of the reagent.xml file implies the use of a WinRE.wim that seemingly you do not have, maybe from it's contents one could try to understand what has "gone wrong": http://www.terabyteunlimited.com/kb/article.php?id=587 jaclaz

Well, historically that is a (mostly unuseful) SAMPLE code for a very limited in features (and with more than one hiccups) RAMDISK driver. JFYI many years ago Christian Ghijselinck took that code and little by little made it into an actually working driver, however nowadays we have much better RAMDISKs around, here is a list of them: http://reboot.pro/topic/1507-ramdisk-and-filedisk-drivers/ more than a few of them are Open Source, so if you are into making a new driver you can have some inspiration. jaclaz

Actually the intended meaning was "While I am actually pretty sure that I could use Clonezilla succesfully for doing that, I pretty much doubt that you will be able to do the same" because I believe that what is done "automagically" in Clonezilla (unless you image the WHOLE disk) is not enough to later perform a "bare metal" recovery. jaclaz

Not really-really. A "Recovery Drive" which has install.swm's is a (poor) replacement for the actual install DVD. If you prefer, there are two kinds of Windows 8/8.1 "Recovery Drives" one which contains only the boot.wim, and one that contains both the boot.wim and the install.wim (or the "partial" install.swm files). It seems that here we have a "third case", with install.swm's and no boot.img @marcusmx5 And what happens if you select the first one? Which EXACT make/model of tablet is it? Each OEM may have introduced one or more changes to the way the "recovery drive" (or the "recovery partition") are created and or should be restored, as an example (related to a Windows 7) Toshiba specifically needed a couple tricks: http://www.msfn.org/board/topic/155075-make-a-recovery-disc-from-a-toshiba-recovery-partition/ jaclaz

Well, You Keep Using That Word, I Do Not Think It Means What You Think It Means ... http://knowyourmeme.com/memes/you-keep-using-that-word-i-do-not-think-it-means-what-you-think-it-means http://homepage.ntlworld.com/jonathan.deboynepollard/FGA/legacy-is-not-a-pejorative.html What I tried to tell you is that while most, but not all, tools are (if you like bells and whistles) working fine, those are the essential things they need to do, while any given tool may fail to get all the relevant data or get too much of them, or fail to restore in the case of a "real" full recovery (i.e. a "bare metal" one). Another thing that you should be aware of is the possible issue with the format in which the data is saved. Many tools use their own "proprietary" format for the DATA, which implies that if - for any reason - the program itself fails in the recovery you cannot use another program instead. As an example, for your intended use, one of the (nice) "features" of Clonezilla, that of saving ONLY used blocks for known filesystems, and then further compress it with gzip, thus providing very compact "images" makes it not so much suitable as it is possible that a recovery partition includes some files that you need to access without restoring the image or some "checks" in unmapped areas, or *whatever*. As well - and as said - a disk may well contain custom MBR+hidden sectors boot code that a tool like Clonezilla may (or may not) save/include in the image. What I would do if I were you would be to do a "real" dd clone of the whole disk, test the clone to be working, then image from the clone through Clonezilla (or other program under test) the parts that you believe are needed to do a "bare metal" recovery, then zap the cloned disk and attempt the bare metal recovery through the tool of choice. If it works, good . If it doesn't, place back the original disk and repeat the test with another tool. jaclaz

Well, if not necessarily, it would be at least "logically". Let's say (hypothetical), that I am a prince of Evil , and I remotely hack PC's (for fun and profit) . I have before me (fiction): 25% of all PC's in the world connected to the internet insecure windows XP systems <- please read as "easy to hack/intrude" 50% of all PC's in the world connected to the internet impenetrable (I told you it was a fiction, didn't I?) Windows 7 systems <- please read as "very difficult to hack/intrude" 12.5% of all PC's in the world connected to the internet even more impenetrable (I told you twice it was a fiction, didn't I?) Windows 8/8.1 systems <- please read as "very, very, really very difficult to hack/intrude" 12.5% (the rest) fractioned in a myriad of less common OS's (that independently from them being more or less secure do not provide a vast enough target)What would I start with? Let me think .... So the logical conclusion is that either XP systems are not so much less secure than later OS's or that Evil princes of hacking do not exist (and they were invented by security firms), of the two possibilities I choose the first one as more probable. There is a third possibility, the Evil princes of hacking do exist, but the security firms are so d@mn good at their work that they succeed in blocking all (or the largest part of) their evil actions before and outside any security measure the OS may provide (which all in all gives the same net result in practice of being not such a difference in security between XP and later MS OS). jaclaz

No, you haven't or they were not "legacy" at all . Basically a partition is (the name says it all) an area of the hard disk, that has a beginning and an end address. These start and end address are recorded in an indexing structure that is a partition table (which is in the MBR or in the first few sectors after it for GPT). So, if you need a part of a whole, you just copy that part and (advised, but not strictly necessary in all scenario's) the information telling you where that part is located (in order to re-collocate it in the same position when restoring or recovery). All you really need is a program capable of copying the sectors from one device to a file, i.e. the traditional (if you want "legacy") UNIX program dd (or any port, derivative or tool having the same capabilities, under Windows I personally use dsfo/dsfi) and *any* program capable of interpreting the indexing data (the partition table). Everything else or "more" that dd (+ a partition table viewer) are (often "nice" but also sometimes "not perfect") feature-filled "overstructures" (user friendly and what not) around these functionalities: understand which and where is the area (i.e. be capable of reading and interpreting the information) copy it understand where it is needed to copy it backMore than the original partition table (which may or may not be relevant, as the information for the partition is anyway in the volume's filesystem BPB ) in the case of the MBR making a backup of it is often VITAL (in the case of these PC's with "recovery partitions") because besides the DATA (i.e. the partition table and the Disk Signature) the CODE contained in it may be "custom" and not easily replaceable, see as an example: http://www.msfn.org/board/topic/131620-hp-notebook-the-recovery-partition-could-not-be-found/ As well some of the so-called "hidden sectors" in a MBR style disk may contain part of the booting code. So the Rule of the thumb is: make a copy of the disk from sector 0 (the MBR) up to the beginning of first partition/volumemake a copy of the desired partition/volumewith these data (in both MBR or GPT cases) you have all you need to restore. jaclaz

There are a couple typos here: FOR /F "tokens=2 %delims==" %A should be: FOR /F "tokens=2 delims==" %%A The snippets posted are not "tested and verified" batches, they are examples/ideas and should be double checked, particularly for (stupid, my bad ) typos. jaclaz

Which translates nicely, in my perverted mind into: since an interesting, new, very valid and proved to be working flawlessly technology conflicts with an old, wrong, stupid update method, the former is removed and the latter (due to some of it's inherent shortcomings and more generally to the lack of a proper method in creating the updates) is dumbed down (if possible at all) and made more complex and error prone for the user. Fascinating as my friend Mr. Spock would define it. Compliments to the MS "Internationalizing Team" I mean jaclaz