TrueImage 2011: Odd partition restore behavior

SAbboushi
SAbboushi used Ask the Experts™
on
Confused...

W7 64 bit.
Partitions in order:
System (primary)
OS (primary)
Extended partition next with 3 logical partitions and a large block of unallocated space at the end

I made a full disk backup (TI2011 latest build as of yesterday).
I tried to restore the OS partition.  After it started, I cancelled.  This left an unallocated block between the System and Extended partitions.
I started to restore the OS partition again.  I selected the unallocated space between the System and Extended partition.

Surprise: TI shrunk the Extended partition so that the unallocated space was now OUTSIDE of the extended partition.  The restored partition was placed AFTER the end of the extended partition:
System
Unallocated (where OS used to be)
Extended with 3 logical partitions and NO unallocated space
OS (primary)
Unallocated

I tried it again - selecting the ORIGINAL unallocated space... but TI used more of the unallocated space at the end of the drive.  And I tried it again trying to figure out what was going on, resulting in:

System (primary)
Unallocated (where OS used to be)
Extended with 3 logical partitions and NO unallocated space
OS (primary)
OS (primary)
OS (primary)
Unallocated

Booting into W7 (I needed to repair first), Disk Management shows me that there are now 4 primary partitions AND and an extended partition - which is surprising to me since the drive is using MBR and formatted with simple/Basic partitions (i.e. maximum should be 3 primary and 1 extended in my case)!!

OK - now the remaining unallocated space at the end of the drive is LESS than the size of the OS partition.  SO I tried again to restore.  I saw that TI wanted to restore as a logical partition, so I selected "Primary" as the type instead.  However, when I went back into W7 after the restore, I see that the Extended partition now included the OS partition that was just restored?! (i.e. the extended partition had grown to include the unallocated space to the LEFT of the extended partition)

System (primary)
Extended with 4 logical partitions and NO unallocated space; the 1st of the logical partitions is the most recent OS restore.
OS (primary)
OS (primary)
OS (primary)
Unallocated

HOW do I control what TI does with the partitions?
e.g.
1) How do I get TI to use the unallocated block that I CHOSE instead of using another unallocated block without even advising me?
2) How do I prevent TI from shrinking an extended partition?
3) How do I prevent TI from GROWING the extended partition to include unallocated space?  How do I get TI to restore a partition as PRIMARY instead of logical when I tell it to do so.
4) How do I get TI to do the things it did when I WANT it to do it (e.g. to shrink or grow an extended partition)
In addition to TI doing something I found unexpected, I am distressed that there was no ACKNOWLEDGMENT that TI was doing any of these things.

I sure hope someone can help me understand this - I don't seem to be finding posts / articles that address what I just experienced.

Thanks -
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
noxchoIT Product Manager
Top Expert 2009

Commented:
Actually interrupting restore process is not the best idea. Restore process where you restore to blank HDD first creates partition scheme in MBR and marks primary partition slot busy. And even if you see that you have unallocated space in MBR the record tells - this place is reserved so use the next vacant slot for new partition. That's why you do jump to new position even if you (as it seems to you) select the right place.
The good and solid approach is delete all partitions from HDD then let the restore wizard restore from partitions using information from backup image. In backup image normally is stored the info about partition structure on your HDD.
If you want to restore only separate partitions from backup then select only those partitions that you want restore. Do not interrupt the restore in half way.

Author

Commented:
>> Actually interrupting restore process is not the best idea
OK - I'll accept that this may have initiated the problem

>> And even if you see that you have unallocated space in MBR the record tells - this place is reserved so use the next vacant slot for new partition
Hmmm... sounds good, but then do you have any thoughts on why TI used this space for a restore AFTER the other unallocated space had been used up?

Anyone have any thoughts on the rest of my questions?
THanks
noxchoIT Product Manager
Top Expert 2009

Commented:
1) How do I get TI to use the unallocated block that I CHOSE instead of using another unallocated block without even advising me?
-Erase the HDD and perform restore partition by partition without interrupting it.
2) How do I prevent TI from shrinking an extended partition?
- See answer to first question. If the MBR has the partitioning scheme that says you have primary partitions then extended one - then you cannot do anything than erasing first two sectors of HDD. Start erase and interrupt it as well.
3) How do I prevent TI from GROWING the extended partition to include unallocated space?  How do I get TI to restore a partition as PRIMARY instead of logical when I tell it to do so.
- If you want the primary to be restored primary show the wizard the unallocated space on healthy drive (not half applied operations and dirty MBR).
4) How do I get TI to do the things it did when I WANT it to do it (e.g. to shrink or grow an extended partition)
- Only during restore process wizard steps. If you interrupt the process and then play again with restore to occupied places you will get to the problems you met now.
How to Generate Services Revenue the Easiest Way

This Tuesday! Learn key insights about modern cyber protection services & gain practical strategies to skyrocket business:

- What it takes to build a cloud service portfolio
- How to determine which services will help your unique business grow
- Various use-cases and examples

Author

Commented:
I found this link that suggests I cannot rely upon W7 Disk Management as always reporting CORRECTLY, e.g. how I could have more than 3 primary partitions with 1 extended partition: http://www.multibooters.co.uk/quirks.html#several_primaries

Any confirmations on this?  Anyone else have thoughts on my questions?
noxchoIT Product Manager
Top Expert 2009

Commented:
Yes. I confirm this. As I told already you can rely on MBR content only.

Author

Commented:

Thanks - much appreciated.  I think I will try the steps again by first deleting the OS partition and then restore to the unallocated space released by the OS delete; if everything works fine this time, this will confirm that it was the cancelled restore that left my disk in an abnormal state resulting in future abnormal results.

Me, I would expect that since TI gives me the option to cancel that it will "clean up" after itself and not leave my partition info in an abnormal state.  I would expect this type of "abnormality" if I did a hard boot during the restore process; I would NOT expect it from an orderly "cancel" where TI has the opportunity to return the sectors occupied by the original partition back to unallocated space (as it SEEMED to have done).

If my expectation is unreasonable, I would be grateful for an explanation of why.  Thanks again
noxchoIT Product Manager
Top Expert 2009

Commented:
Cleaning up the disk after cancelling the operation is pointless. If you start restore mistakenly to wrong partition and then realize you have there very important data - you would rush into cancel. Agree?
Then try recovery software to browse this space on HDD. But if cancelling erases the drive - you are left with no chance for recovery. Thus cancel leaves the state of the HDD as it is at the moment you press Cancel.
This is just common approach for all backup & recovery software, not only Acronis. I am familiar with almost 10 backup and recovery software tools from different vendors and they work the same way.

Author

Commented:
Thanks - good explanation.  

I want to differentiate between "erasing the drive after cancel" vs. "making sure the partition table is valid after cancel".

I understand your point that if I cancel, I do not want TI to overwrite any further sectors.

However, after cancel, let's consider these 2 scenarios:
1) I mistakenly selected the wrong target partition for the restore (your scenario), so I cancel with the objective to recover whatever was not overwritten
2) After starting the restore, I realized there was something I wanted to examine and didn't want to wait for the restore to complete first (my scenario) .  So I cancelled, booted from a flash drive, and examined what I wanted.

With either of these scenarios, the NEXT step would be to do something with the unallocated space left behind as a result of the cancel.  I am having difficulty with the concept of "if you cancel a restore operation in TI, your MBR is

You characterized the MBR as being "dirty" after the cancel.  What exactly do you mean by that?  I believe there are 16 bytes allocated to each partition slot to define a partition.  I assume you mean that those 16 bytes specific to the partition I was restoring (but cancelled midway) are somehow dirty?  If that is the case, my argument is that TI is responsible for making sure that is NOT the case by setting those 16 bytes to reflect an UNUSED partition slot (which I have no reason to believe it did NOT do other than posters who are suggesting otherwise)

Also, maybe you can clarify for me: is it TI or is it the BIOS that is responsible for managing the MBR partition table?

THanks - I appreciate your efforts in helping me to understand what is going on
noxchoIT Product Manager
Top Expert 2009

Commented:
By dirty I meant - not empty, having records about existing partition(s) on the drive. The process of building new partition on HDD (lets consider the drive is blank and new) goes the following way:
1)Create record about new partition in MBR (erase existing record and replace it new one)
2)Create partition (specify its start and end)
3)Format partition into file system
4)Copy data to partition and update partition boot record.
This is the order of restore.
So if you cancel the operation on first step then you have record in MBR (dirty) but no partition yet.
If you cancel it on second step then you have unformatted partition on HDD that you cannot do anything with than deleting
If you cancel it on third step then you have blank partition with no data.

Author

Commented:
>> So if you cancel the operation on first step then you have record in MBR (dirty) but no partition yet.
OK - but how does it make sense for TI to leave such useless info in that MBR partition slot instead of setting the slot to "unused" after a cancel??  Practically speaking, I imagine that creating a 16 byte record (the MBR slot for the partition) will happen faster than I can press the cancel button?

>> If you cancel it on second step then you have unformatted partition on HDD that you cannot do anything with than deleting
I'm OK with that - the partition has been defined but I would have a corrupt or no file system requiring a format/reformat.  MBR slot would still contain valid data, wouldn't it?

>> If you cancel it on third step then you have blank partition with no data.
I'm OK with that too - again, MBR slot would still contain valid data, wouldn't it?

So it seems to me that the only case that would result in bad data in an MBR slot is for the first scenario... and again, if writing 16 bytes to the slot is the FIRST step, then what are the chances that a "cancel" can interrupt the writing before it completes?

Hopefully my response has made it clear where maybe I have some misunderstandings and you can correct me.  Thanks-
IT Product Manager
Top Expert 2009
Commented:
OK - but how does it make sense for TI to leave such useless info in that MBR partition slot instead of setting the slot to "unused" after a cancel??  Practically speaking, I imagine that creating a 16 byte record (the MBR slot for the partition) will happen faster than I can press the cancel button?
This record in MBR does not mean that data on HDD surface is overwritten yet. Thus the approach is - leave as is at the moment of cancel.


I'm OK with that - the partition has been defined but I would have a corrupt or no file system requiring a format/reformat.  MBR slot would still contain valid data, wouldn't it?
The fact that file system is corrupt does not mean that your data is totally gone. Special data recovery services could recover it still. Or even some software like GetDataBack. Depends on the fact if the overwriting process did start or not. It starts normally at last step. So the FS would be corrupt for system only.


I'm OK with that too - again, MBR slot would still contain valid data, wouldn't it?
MBR contains only reference to the place on HDD where partition must be. This is not a valuable data really. Data is on HDD not in its start.

So it seems to me that the only case that would result in bad data in an MBR slot is for the first scenario... and again, if writing 16 bytes to the slot is the FIRST step, then what are the chances that a "cancel" can interrupt the writing before it completes?
Chances are actually high. Normally it works as I said. Record in MBR is created - then in defined borders partition is created and formatted then data is copied there. The last step is most dangerous.

Author

Commented:
THanks again for all your help.

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial