Solved Benefits of GPT vs MBR Primary and MBR Logical Partitions ?

[SIW2]

This whole efi/gpt thing certainly adds complexity and introduces limitations for the end user. They don't get much choice, pretty much all the new laptops the average Joe buys come like that now.

 

[jumanji]

.......and therefore it will be wiser for anyone with GPT drives to save all the important sectors using Bootice so that one can restore the drive to the original condition in case partition related problems arise making the drive inaccessible.

The screenshot below shows a GPT drive with three partitions.



In this case as an example, I would save LBA0(sector 0), LBA1, LBA2, LBA34, and then the three NTFS volume boot records LBA 262178, LBA 366634632 and LBA 732818343.

During the restoration process I would start with restoring the backups that are supposed to exist in the last sector (LBA n) and LBA n-32. These might have gotten corrupted and need to be restored first.

So I would restore the last sector n with the GPT header - saved LBA1, and then sector n-32 with the GPT Table - saved LBA2.

EDIT: The following figure shows the partition structure of a GPT drive.



That should give an idea what all sectors one should backup and save. That would depend upon the number of GPT partitions one has. While restoring the saved sectors to the backup area in the drive the following progression should be maintained.

LBA n (the last sector) with saved LBA 1 and then on LBA n-32 with saved LBA 2, LBA n-31 with saved LBA 3, LBA n-30 with saved LBA4 .......... upto LBA n-1 with saved LBA 33

Then LBA 0, 1, 2,.......to 34 and other VBRs with the respective saves.

In most cases, one should be back in business and be able to access the drives as before without the necessity of trying any data recovery software the efficacy of which in dealing with GPT drives, in my view, is uncertain as on date.

On how to save and restore the sectors using Bootice, http://www.sevenforums.com/hardware-devices/310295-lost-partitions.html#post2584426

Update: Amplified in post#29 to be read in conjunction with.

 

[SIW2]

Far,far,far too complicated for the average user.

 

[jumanji]

I am an average user , honestly. ( looks, reads very complicated but once you do it you know you can do it and if I can do it then anyone can.)

 

[milindsmart]

I think this is the KB article in question : Disk drive numbers may not correspond to the SATA channel numbers when you install Windows on a computer that has multiple SATA or RAID disks
My goodness... What an atrocious error! This is absolutely mind-blowing in stupidity. Drive numbering according to.... order of enumeration.
Even with this brain-dead manner of drive numbering, how could it get confused with ITS own numbering in a SINGLE boot!?

And yeah, the tool is apparently buggy, if it can't find the simple truth that this disk was GPT.
But wait, could you not tell it that this was a GPT disk, and so not to try to interpret it as MBR?
Perhaps your classification of partition managers as Beta is right... It's so much easier to recover from GPT, I can only blame the tool.Reassigning GUID is trivial. simply replacing with a protective MBR would have let any other tool recover the backup partition table.

Windows is to blame for getting confused, but Lazesoft is to blame for not recognizing the basic GPT layout. There is NO 3mb partition connected with GPT at all. Wonder what it scanned...

 

[milindsmart]

SIW2, What do you think of this idea of mine?
gpt - Out of Band Disk Partitioning and Tables - Super User

Would like some super-expert opinions

 

[mjf]

As I've said the vast majority of users don't need a GPT boot disk (really useful on a 250GB SSD!). However, as SIW2 says it's pretty much standard on most bought PCs so if you are interested in computers it is probably worth understanding the GPT standard.

 

[alan10]

I think this is the KB article in question : Disk drive numbers may not correspond to the SATA channel numbers when you install Windows on a computer that has multiple SATA or RAID disks
My goodness... What an atrocious error! This is absolutely mind-blowing in stupidity. Drive numbering according to.... order of enumeration.
Even with this brain-dead manner of drive numbering, how could it get confused with ITS own numbering in a SINGLE boot!?

I totally agree.

Windows is to blame for getting confused, but Lazesoft is to blame for not recognizing the basic GPT layout. There is NO 3mb partition connected with GPT at all. Wonder what it scanned...

I remembered wrongly, and Lazesoft made absolutely no such mistake.

I actually tried many third party tools and they were all defeated by Windows insisting that the GPT DIsk had to retain the MBR style of Disk Identity.
Perhaps I could have booted into a LINUX Recovery tool and fixed the problem,
but I thought it more likely that Learning to use Linux tools would result in me placing everything in jeopardy.

It was the commercial MINIDISK DATA RECOVERY which showed me 3 MB things which I assumed were remnants of GPT links.

Lazesoft actually showed me 1.41 MB EFISECTORS, which I assume are your GPT linked list.
Lazesoft was NOT being used to recover partitions but was digging through UNALLOCATED SPACE in that region of the disc, and it did the job perfectly.
It rescued over 400 GB of Macrium Partition Image backup files,
and Macrium has validated all the internal checksums of each file,
THEREFORE all the fragments were correctly assembled (the files were fragmented when first created).

The only alternative data recovery that recovered each file with the correct number of bytes,
probably assumed that no file was fragmented,and simply located the start sector for each file and then read a contiguous set of sectors, actually composing a "recovered file" which consisted of fragments of many different files. Macrium tested its recovery attempts and each one failed hash checksum validation.

Lazesoft now has a recently enhanced capability to repair in a few minutes all the damage done by Windows when Windows "converted" the GPT disk into a trashed MBR disk,
BUT I will only test this once I am satisfied that I have NOT RECOVERED ANY FILES that I actually deleted BEFORE Windows has trashed the system.

Regards
Alan

 

[jumanji]

Having gone through all the posts in this thread this morning and having noted that Windows can create problems by not enumerating the disks in the same order with each startup and presenting a different picture to the unwary user, I am more than ever convinced that users with GPT drives will be better off saving the important sectors in all the drives they use - whether MBR or GPT.

In the case of MBR disks the screenshot below shows the typical LBAs one would save for later restoration. (This disk has three primary and three logical drives.)



I have since amplified my post #22 by including a figure showing the GPT structure. I repeat it here with some colours added.



On a GPT drive if one has 128 partitions then one may have to save roughly 38+128(VBR) LBAs.

But considering that nobody is likely to have more than 12 partitions the number of LBA saves get reduced to 6+12(VBR) LBAs. Each sector save is only 512bytes.

And saving the LBAs with Bootice is not at all a difficult task. One just selects the drive, clicks on Sector Edit, view the partition details, select the sectors one by one and click on the save to file icon. Similar process for restoring. ( In the case of restoring the sectors on the system drive that has gone awry one may have to perhaps run Bootice from a WinPE disk or remove the HDD and place it as a second drive on another system and run Bootice). By default Bootice saves the files in the same folder in which it resides.

The only problem an average user will face is what goes where when restoring. The GPT structure figure coupled with the text statements should come in handy for that. Does not take much time to figure out and understand.

If anyone who wants to undertake this exercise face any problem, we are always here to walk you through.

 

[mjf]

The enumeration sata port issue has been raised here many times before. I lean towards a BIOS explanation as proposed here:
http://www.sevenforums.com/installa...rt-disk-management-vs-bios-2.html#post1519407
Now does a UEFI BIOS solve this --- who knows. But for all practical purposes it may not matter.

 

[Rockfella]

Which one is faster?I tried both and felt MBR to be slightly faster than GPT. I don't have crazy drives hence don't really need GPT but my motherboard supports it so went ahead and tried it. It feels slightly slower than MBR. Anyone with any experience using both?

 

[ignatzatsonic]

Which one is faster?I tried both and felt MBR to be slightly faster than GPT. I don't have crazy drives hence don't really need GPT but my motherboard supports it so went ahead and tried it. It feels slightly slower than MBR. Anyone with any experience using both?

Never noticed any difference, but I didn't break out a stopwatch or run benchmarks.

The main reason I'd use GPT is to be able to use all of a large drive and that's much more important to me than whatever miniscule speed advantage MBR might have---if any.

 

[milindsmart]

I'm very surprised that you found a speed difference.... Do you mean boot time? Other than that there should be absolutely no difference.

 

[GeneO]

I am an average user , honestly. ( looks, reads very complicated but once you do it you know you can do it and if I can do it then anyone can.)

Why is this any different than MBR? GPT has redundant structures that make it less likely the disk boot sectors will be unrecoverably corrupted than with MBR and your are making it sound worse.

Why would you need to save both the primary and backup sectors- you need only save one or the other.

If the data is important, make a backup image or have a copy on another drive. Being able to fix a drive's boot sectors only gives you a sense of false security IMO. If these sectors get corrupted, it is likely your data is ass well.

 

[Rockfella]

I just felt like that but I was wrong. Some youtube videos suggest GPT is faster so I am using GPT now.

Which one is faster?I tried both and felt MBR to be slightly faster than GPT. I don't have crazy drives hence don't really need GPT but my motherboard supports it so went ahead and tried it. It feels slightly slower than MBR. Anyone with any experience using both?

yeah boot/reboot only. Although I didn't ran a stop watch. Just went by this review:
https://www.youtube.com/watch?v=dQIE5tV2LAs

I'm very surprised that you found a speed difference.... Do you mean boot time? Other than that there should be absolutely no difference.

 

[alan10]

With all due respect, data backups are NOT RELEVANT to the topic in hand,
which is concerned with the relevant merits of GPT and MBR style Disks.

Post #7 indicates how useless GPT can be as a backup because a backup that fails is USELESS as a backup.

I received a Windows 7 Desktop already configured with System C: installed on an MBR Disk because the hardware/BIOS lacked the ability to boot from GPT.

My main use for the GPT disk was to create and maintain on-line Macrium Reflect image backups of C:\,
and to frequently duplicate those backups to USB and eSATA connected disks.

Life is too short to make a backup of every backup of EACH backup.
So I do not waste my time,

One morning Windows woke up stupid,
and decided that the GPT disk had the disk identity of my MBR disk, and put my MBR disk (C:\) offline.

I was able to eventually give the original MBR disk a new identity to replace that which had been stolen,
and put the imitation MBR disk off-line.
Windows does NOT allow me to correct the disk ID of the original GPT DISC, and that puts ALL on-line backups out of bounds.

I needed Lazesoft Data Recovery to retrieve those backups to off-line storage.

I have seen no reason to defer changing my DEAD-OFFLINE Disk to MBR,
and this should avoid any repetition of the same Windows stupidity.

Regards
Alan

 

[GeneO]

From what I have read, it wasn't as simple as your system woke up stupid. You booted to a SysLinux flash drive, it crashed, and when you rebooted you found that the GPT disk was no longer recognized as such. That the flash boot failed is an indication it was doing something stupid. If I had to guess, it would be that the flash boot saw the disk as a damaged MBR and overwrote the partition type of 0xEE and tried to repair it, turning it into a regular MBR. The partition type had to have been changed, and when it did, it was recognized as a valid MBR partition. You may have tried this, but I would beet that if you changed the partition type of the MBR back to 0xEE, you would have recovered your gpt disk.

 

[jumanji]

I am an average user , honestly. ( looks, reads very complicated but once you do it you know you can do it and if I can do it then anyone can.)

Why is this any different than MBR? GPT has redundant structures that make it less likely the disk boot sectors will be unrecoverably corrupted than with MBR and your are making it sound worse.....

''GPT was designed with improved robustness in mind compared to the older MBR partition table;GPT includes stored CRC values to help utilities spot problems and redundancy to help recover from them. Neither of these features makes GPT completely immune to errors, though; they just help utilities to detect and recover from those problems" - Rod Smith, Author of GPT fdisk- a set of text-mode partitioning tools for Linux, FreeBSD, Mac OS X, and Windows that can also repair/try to repair a broken GPT.

The same author: " An Ounce of Prevention: Before you get into trouble, it's worth taking preventive measures: Back up your partition tables! If you take the simple precaution of backing up your GPT data, and keeping this backup current with any changes you make, you should find data recovery much easier."

This learned author sounds worse than me, an average user !

 

[alan10]

From what I have read, it wasn't as simple as your system woke up stupid. You booted to a SysLinux flash drive, it crashed, and when you rebooted you found that the GPT disk was no longer recognized as such. That the flash boot failed is an indication it was doing something stupid. If I had to guess, it would be that the flash boot saw the disk as a damaged MBR and overwrote the partition type of 0xEE and tried to repair it, turning it into a regular MBR. The partition type had to have been changed, and when it did, it was recognized as a valid MBR partition. You may have tried this, but I would beet that if you changed the partition type of the MBR back to 0xEE, you would have recovered your gpt disk.

Where were you when I asked years ago how to turn my disk back to GPT ! ! !
ALL INTELLECTUAL GIANTS AT THAT TIME FAILED TO SOLVE THE PROBLEM.

Please explain how and why Windows would decide that it was appropriate to IGNORE all the good disks,
and deliberately choose to not only re-write a long and complex GPT GUID as a simple and short 8 digit number,
but to ALSO ARBITRARILY USE THE SAME 8 DIGIT NUMBER EMPLOYED by my system C:\ MBR SSD Disk,
and THEN PUT MY SSD OFF-LINE,
and try to boot my trashed GPT disk as C:\ ! ! !

Software such as Macrium Reflect and Hard Disk Analyzer,
AND ALL THEIR COMPETITORS,
were dependent upon disk numbers, and that was good with IDE.
All partition image users had problems once SATA came along,
and Macrium Reflect were the first to overcome the new idiosyncrasies with Windows.

Obviously Windows itself is fundamentally flawed and cannot start up in a predictable controlled fashion.

Yes, I have dabbled with SysLinux
BUT NO -SYSLINUX WAS NOT PART OF THE PROBLEM,
and I cannot find where in this topic you had cause to think it was.

WINDOWS IS STUPID.
It does NOT number SATA disks according to how they are physically connected but in "order of enumeration",
which is MickyMouse speak for "a race hazard we cannot handle - just deal with it."

My SSD is always enumerated last, as number 3,
BUT ONLY when I first switch on the P.C. at the start of the day.
If I ever REBOOT then the SSD is NORMALLY either number 1 or 2.
Further evidence of a STUPID WINDOWS RACE HAZARD.

Regards
Alan

 

[GeneO]

From what I have read, it wasn't as simple as your system woke up stupid. You booted to a SysLinux flash drive, it crashed, and when you rebooted you found that the GPT disk was no longer recognized as such. That the flash boot failed is an indication it was doing something stupid. If I had to guess, it would be that the flash boot saw the disk as a damaged MBR and overwrote the partition type of 0xEE and tried to repair it, turning it into a regular MBR. The partition type had to have been changed, and when it did, it was recognized as a valid MBR partition. You may have tried this, but I would beet that if you changed the partition type of the MBR back to 0xEE, you would have recovered your gpt disk.

Where were you when I asked years ago how to turn my disk back to GPT ! ! !
ALL INTELLECTUAL GIANTS AT THAT TIME FAILED TO SOLVE THE PROBLEM.

Yes, I have dabbled with SysLinux
BUT NO -SYSLINUX WAS NOT PART OF THE PROBLEM,
and I cannot find where in this topic you had cause to think it was.

DISKPART UNIQUEID Error. Windows reports my GPT Disk as MBR. How do I fix it ?

I found this post (I assume to be yours) looking for failed GPT as I couldn't imagine the scenario you presented. It sounds to me like the smoking gun is the failed Linux boot. Though I suppose it is possible for something to have happened before the Linux boot that made it crash.

WINDOWS IS STUPID.
It does NOT number SATA disks according to how they are physically connected but in "order of enumeration",
which is MickyMouse speak for "a race hazard we cannot handle - just deal with it."

My SSD is always enumerated last, as number 3,
BUT ONLY when I first switch on the P.C. at the start of the day.
If I ever REBOOT then the SSD is NORMALLY either number 1 or 2.
Further evidence of a STUPID WINDOWS RACE HAZARD.

Regards
Alan

Windows doesn't use the enumeration to address drives, it uses the disk's signature which is an ID written on the disk and read at boot time. I don't know if windows can do anything about the order that drives are enumerated or if that is a BIOS thing, but it is what it is. It certainly is a pain to diagnose problems I agree. You also have to be careful with programs like DISKPART - that your are operating on the disk that you think you are - or you can clobber a disk that way.