Vista: Are there issues with the disk ID?

Hi MudCrab,

I'm sure my Vista had been previously repaired. It's been through the wars of multiple testing. I know it's been restored with TI many times and also with Vista's Backup solution at least once. I have VistaBootPro installed and it shows some interesting information. On the screen that shows BCD Settings I see thee different sections for Windows Boot Loader. Two of the three are labled (recovery) in the desription. You need to check the "All" radio button to see them. This led me to the registry at HKLM\BCD00000000. You can drill down to see Objects. These Objects seem to represent identifiers, resume objects and resume from hibernation objects amoung others.

I did a fresh install of Vista on a wiped drive. On first boot I looked in the registry and found there was no BCD hive. The system then downloaded some updates and wanted to reboot. I let it. When it came back up, I looked in the registry again. Now there was a BCD hive. It showed 11 Objects.

I did a ful disk backup to test your senario. I got the same result as you. Restoring the system partition and the MBR (not the entire disk) on top of the existing Vista partition produced a winload.exe missing error. Restoring the entire disk worked. This showed that my previous results were better because the sytem had been repaired before the image was made.

Next, I created a logical data partition on the fresh Vista disk for data. I backed up just the system partition and the MBR with both TI10 and Paragon Hard Disk Manager 8.5. There was no way to get TI to restore the system partition on top of the existing partition without doing a Vista Statup Repair. Paragon Hard disk Manager 8.5 restored just the system partition without a hitch. I noticed the last message it gave during the restore process was something about updating the Boot Configuration Data registry hive. Vista booted up like nothing had happened. I looked at the BCD hive and I only found 11 objects. VistaBootPro still only showed one entry under Windows Boot Loader.

Then I went back and restored with TI and ran a Vista "Startup Repair". Vista booted up and found new hardware and required a reboot. After I rebooted, I looked at the BCD registry hive. Now there were 13 objects. VistaBootPro now showed two Windows Boot Loader entries. The one labled as (current) also showed as (recovered). Aparently, these two extra objects are helping TI handle a restore after you take a new image.

These new results are somewhat disappointing. Paragon HDM is looking pretty good. It also runs nicely under BartPE in case the Linux Recovery CD doesn't support your hardware.

 

mustang,

Thanks for the further testing. At least the results are consistent.

Did you notice when you used Paragon HDM if it "moved" the Vista partition? In other words, the "recovered" partition is not starting and ending on the exact same sectors. TI does move the partition and I think that's what causes the problem. Why it moves it beyond me. I would think it would be best to restore the exact partition back to the exact same place with the exact same ID. Instead, TI deletes the partition, creates a new one in a different place and then restores. This is why the restoring of a repaired partition works (in my opinion, anyway). TI has already moved the partition and then it puts it back in the same moved placed that the repair fixed it to work at.

 

I have an interesting update.

I did another restore of the two partition image that was created wth TI10. This time I restored it on top of the repaired system. As you will recall, it was only a partition (not a full disk) image. This time I restored the image with TI9.1 Workstation build 3887. Guess what. It booted without a repair! Hopefully, this says something good about the next build of TI10.

I'm going to start all over and do a fresh install of Vista to a wiped drive. This time I'll image and restore it using Workstation. That will eliminate the possibility that retoring the system partition on top of a repaired partition had something to do with my success restoring with Workstation.

Be back soon with the results. I'll also see if I see any signs of the partition moving. Was your system partition the first one on the disk or was there a diagnostic partition in front of it?

 

I have some good news and some confusing news.

I wanted to test just restoring a system partition using TI9.1 Workstation on a multi-partition disk. This time I started with a wiped disk. I used Paragon Hard Disk Manager 8.5 to create three partitions on the disk. The first was a 54.8 MB FAT16 hidden partition to simulate an OEM diagnostic partition. The second was a 64 GB NTFS primary partition for Vista. The third was approximately a 50 GB logical NTFS partition for data. I did a fresh install of Vista and backed up only the system partition with TI Workstation.

The good news is that restoring just the system partition on top of the existing system partition worked perfectly. No Startup Repair was needed. I didn't even restore the MBR.

The confusing news is that doing the same restore with TI10 also worked. This peaked my curiosity. I then did a backup of only the system partition with TI10. Again, restoring it on top with TI10 worked. I'm beginning to think this whole issue has something to do with how the system partition was created in the first place. In my earlier tests of restoring a fresh Vista system with TI10 that refused to boot, I had let Vista create and format the partition in unallocated space during the install. This time I pre-created the partitions using HDM 8.5.

It would appear that predicting how this will all work on other computers may be difficult.

MudCrab, how did you create your Vista partition? Could you try something similar to what I just did by creating the Vista partition with different methods and see if your restore results change?

 

mustang,

In all cases except for the BootIT NG tests, I used Vista to create the Vista partition. I think that is why the BING Vista partition restored successfully and the others didn't. I agree that it has to do with how the partition is created. On my test system Vista made the "relative sectors" at 2048 on the drive instead of the TI default of 63. This is where TI moves it when it recreates the partition.

All tests were made with Vista Home Premium OEM version installed by booting off the DVD. There were no other partitions except those stated in the tests. In the case of BING, there is a small special BING partition at the beginning of the drive.

I ran the tests with Vista on the first partition. I wonder if it would still create the offset if Vista is the second partition and XP was first. It's possible that Vista only creates this large offset on the first partition.

I only used Vista to create the Vista partition (except for the BING tests) because Acronis Disk Director has some problems creating a partition that works with Vista. I was trying to simulate a clean install of Vista in which case the partition wouldn't already exist.

The partition issue is why I was curious if you could check if the partition moved from when the backup was made to after it was restored. Move = Repair needed. No Move = No Repair. That's what I'm thinking. Now I'm going to have to run some more tests...

 

mustang,

I had time for one test tonight. I wiped the drive and then created one partition for Vista using Disk Director (build 2,160). I checked that the offset was sector 63 (it was). I then installed Vista into that partition (I did reformat the partition with Vista before installing). After Vista was installed and updated I created a backup image of the drive. Then I restored just the partition. It booted up fine, no repair needed.

I still want to test installing Vista to the second partition. I suspect that it will restore without needing a repair.

 

Hi MudCrab,

Now we're getting somewhere.

Good to hear about your latest test. It's also good to hear that DD10 build 2160 pre-created a Vista partition successfully. The serious problems with build 2117 and Vista were the reason I used HDM to pre-create the partitions.

I still need to do one more test and let Vista create a single partition on a wiped drive. If I get the 2048 sector offset vs the 63 sector offset, we have this thing figured out.

 

Okay, I did the fresh install again. The Vista created partition did have the 2048 sector offset.

This gives me confidence again that TI10 will do a good job restoring the system partition only. Most OEM computers will probably not have Vista in the first slot with a 2048 sector offset. Also, anyone who did install Vista in the first partition and let Vista create the partition should have the install disk to run a repair if needed.

I'll be gone for the rest of today. Tomorrow, I'll backup and restote this latest Vista with both TI10 and HDM and see how the restores go.

 

mustang,

I have run two further tests.

Test #1 - Install Vista on the second partition using Vista to create both partitions.

Test #2 - Install Vista on the second partition using DDS (build 2,160) to create both partitions.

In both tests, I wiped the drive before starting over and I had Vista Install format the partition prior to installation.

Test #1 Results - Restore of just the Vista partition required a repair. The partition had been moved by TI. The first partition (unused) had the "normal" Vista offset of 2048, the second also had a higher offset though not as much as the difference between 63 and 2048.

Test #2 Results - Restore of just the Vista partition worked perfectly. No repair was needed and the partition was not moved.

So it seems clear that the problem is that any partition created by Vista (at least during the installation process) will require a repair if restored via TI. Any partition created with DDS (build 2,160) or Paragon HDM 8.5 works fine.

At the present point, to get the most trouble-free use from TI and Vista, create the partition for Vista using something other than Vista (DDS, Paragon HDM, BING, etc.) and then install to that.

I assume that Acronis is aware of this problem and hopefully it will be fixed in the next build. To me it seems like a simple problem and the solution is to just restore the partition back to it's original place (or check the Vista boot sector and find out what the linking sector is and put it there, or go ahead and restore like it does currently and then repair the link so Vista can boot).

 

It’s only ever the first partition on a drive that Vista will change the traditional starting sector. It’s all to do with future large-sector hard drive support.
http://support.microsoft.com/kb/923332

 

Thanks McTavish for that link.

Now that we know that Vista is using a new format to create NTFS partitions, caution should be advised. I think the safest practice would de to do full disk backups. I don't have much time to research the new standards and come up with all the procedures on how they will impact imaging/restores with TI10. It would be nice to here from some of the users of this forum with OEM machines that came with Vista pre-installed. Are you able to restore only the system partition without running into the missing winload.exe error?

In my last post I mentioned that I would try backing and restoring a Vista created partition in the first slot on the drive using HDM 8.5. We already know that restoring only the system partition using TI10 requires a Startup Repair to get Vista working. I wanted to see what would happen using HDM 8.5 in this situation. It turned out to be much worse. After the restore, I had a disk that would not boot in any way. The system froze during the initial post. It got as for as looking at the Vista HDD and reported SMART status was OK. That was it. No way to even boot from a floppy or CD/DVD with the drive attached. I couldn't even enter the BIOS setup. I had to remove the data cable from the drive to get past the initial post. I put the Maxtor diagnostic floppy in and booted. Just as I got past the post, I plugged the data cable back into the drive. The drive was detected and I did a low-level format to make the drive usable again. That was late last night and I was tired. I should have booted from the Vista install disk with the plugin on the fly method to see if a Startup Repair could fix the problem. I may do that later.

To BrianK or RAD,

Is there any feedback in the RAD forum on how Ghost 12 Beta is handling this situation?

 

MudCrab:

Nice work. I've been following your experiments with great interest.

It would also be interesting to see if the DD10 Disk Editor could repair a re-imaged Vista partition. If the partition was moved from starting at sector 63 to starting at sector 2048, could you use the disk editor in DD10 to change the value of "sectors before" in the partition table from 63 to 2048 (or vice-versa)? If this works it would avoid the need to have a Vista DVD to repair the restored partition.

 

McTavish,

Yes, thanks for the link. That explains what Vista is doing. However if you assume that TI is creating standard partitions when it recreates them, then it's not only the first partition that Vista places differently.

Below are pictures of two partitions created with Vista. The second one (on the bottom) is the one with Vista installed.

Installed system
http://www.purviancecs.com/images/acronis/VistaOn2nd_before_restore.jpg

After TI restored the Vista partition
http://www.purviancecs.com/images/acronis/VistaOn2nd_after_restore.jpg

As you can see, TI moved the second partiton. This causes a repair to be needed. I suppose that Vista is creating the second partition using the same "offset" guidelines for new larger drives and TI doesn't recognize it.

Note: The reason the first partition in the "before" picture shows FAT16 is because I hadn't formatting it yet when I took the picture. Since the partition sectors didn't change, I didn't bother redoing the picture.

 

Hi, Mark. Thanks again for all your help building my "rescue tool". I use it all the time (like for doing these tests).

That is an interesting idea to try and position the partition back to it's original place. I'll give it a try and see what happens.

 

An important point has come out of this thread, namely the Visa start sector offset. I also wonder whether this affects DDS 10 Build 2160 in any way?

It's been a whole week now and still no comment from Acronis Support. I will PM them and request that they call in here to advise whether or not their Development Team are aware of this issue.

Regards

Menorcaman

 

Mark,

I looked into using DD to change/move the sectors of a Vista parititon. It doesn't look like it could work. You can change the sectors in the partition list, but that won't change the location on the disk (it will point to the wrong place). Vista places the bootsector data on sector 2048. Between the MBR and there is all blank. When TI restores, it places the bootsector at 63 and continues from there. Sector 2048 now contains data.

Also, trying to slide the partition doesn't work as DD doesn't provide the resolution needed for the procedure (it only moves in increments of 7.813 MB).

Menorcaman,
While Vista seems to install and run okay on a partition created with DD 10 (build 2,160), it obviously does not create a Vista compatible partition in the sense of it being like the one Vista creates. There are also no options or controls to select whether the partition should be Vista or XP NTFS. Even if there were, TI presently would ignore it and restore as current default just like it does with Vista created partitions.

On current hard drives I doubt that this will actually cause any problems, but in the future if someone upgrades from a 500GB drive to a 3TB drive (with large sectors) then TI (or whatever program is used) will have to recreate the partition and format structure to be compatible.

TI must keep track of the sectors the partition is using so it must know where it starts and ends. What I don't understand is why when it recreates it, it doesn't use that information to place it correctly. Even if a resize was performed on the partition during the restore, it shouldn't affect the sector offset. In other words, if a Vista partition is backed up, then a Vista partition should be restored, not a modified version of it.

That being said, TI must do something differently when the whole drive is restored because it keeps the offset. The Vista partition is created just like it was originally. TI just needs to apply this same functionality to restoring single Vista parititons.

 

If Vista originally placed its partition starting at sector 2048, then I assume that when you examine that partition using DD10's Disk Editor you should see the parameter "Hidden Sectors" in the partition table set to the value "2048". Correct?

And if TI restores this partition but starts it at sector 63, then I also assume that the parameter "Hidden Sectors" gets changed to "63". Also correct?

If both of these statements are true and Vista won't boot correctly, then there must be an entry somewhere in the new Boot Configuration Data file that references the location of Vista by absolute sector, or somehow depends on the location of the starting sector (checking the disk id, for example).

I didn't mean to suggest sliding the partition's data around; rather to check for and repair an incorrect "hidden sectors" entry. But perhaps this isn't the case?

On my Vista machine, I had created the partition structure originally with DD10 when it was running XP. Then I did a clean install of Vista, but did not have the Vista installer reformat the partition. Looking at it now, the first primary partition contains Vista and starts at sector 63, and the value "hidden sectors" in the partition table is 63. From what you've found out, if I had let the installer format the partition it likely would have started at sector 2048. And if I were to restore one of my backup images using TI10 the partition should end up in exactly the same place and it should reboot without error (sorry, I can't do that experiment now because the machine is in daily use at work!)

What you all are discovering is very interesting. I had at first thought that Vista made changes to NTFS, but Microsoft says that it didn't change the NTFS filesystem in Vista. Apparently, what they did was to make allowances for the upcoming new large hard disks by aligning the start of the Vista partition on a boundary that is different from past practices (starting at sector 2048 instead of sector 63). It is apparently this change that must be accommodated by programs like DiskDirector, TrueImage, PartitionMagic, etc.

 

Yes

Yes

This seems to be the case. I'm sure that it's just a simple matter of pointing the loader to the correct sector. BING does this by letting you edit the BCD file and just selecting the partition, so it can't be that hard to do.

The hidden sector entry is correct (at 63) after the restore. The boot sector is also on sector 63 so they are in agreement. Apparently, Vista is still looking at sector 2048. I was checking into moving/sliding to find out if I could make DD move the partition so that it started on sector 2048 instead of 63.

This is not correct. I've only gotten the 2048 offset if I let Vista create the partition. Letting Vista format an existing partition (one created with DD, for instance) does not change the offset.

If you were to restore a backup of your Vista partition (that is at the 63 sector offset) then, yes, it should work perfectly. I've done this in my tests and also my P5B computer and there is no problem with it.

 

Excellent. It's starting to make more sense now. Next I really need to learn more about Vista's Boot Configuration Data file.

Thanks for your insights on this topic.

 

Hello everyone,

Thank you for choosing Acronis Disk Backup Software.

We are sorry for the delayed response.

Please be aware that there are several cases in which one should use Windows Vista DVD to perform Startup Repair after image restoration/cloning process. We are aware of this issue and Acronis Development Team is working on it. We are now closer to resolving the problem.

Note that some of the issues were already fixed and we will do our best in order to provide you with a new build of the particular version of Acronis True Image, which contains the fix as soon as possible. I'm sorry, but at the moment the exact time-frame for that is not set yet.

Please accept our apologies for the current inconvenience and once again, thank you for your assistance and understanding.
--
Aleksandr Isakov

 

I think I may have misunderstood something regarding the BCD here.

Two cases where restoration works are a) where a disk restore is carried out and b) where a partition restore is carried out into a partition which has been created by something other than the Vista installation process.

If a restoration is carried out after Vista has created its partition, and TI subsequently restores it back to sector 63 isn't the problem with the fact that the MBR is trying to load the primary partition from sector 2048?

If this is true (and as I say I may have missed something) how does the BCD come into play as it has not even been found yet.

Thoughts please.
Thanks
F.

 

Hello Aleksandr Isakov

Thank you for your contribution to this thread.

I find it somewhat disturbing that it took 8 days and the intervention of a Global Moderator before Acronis could find the time to respond to this very critical issue.

Quote from your post...

I assume that Acronis Engineers are aware that 95%+ of users purchase their machines with Vista preinstalled under an OEM License?

I must also assume that your engineers are aware that the CD's provided by resellers under an OEM License contains a subset of Vista and may not contain the Startup Repair?

I must further assume that your engineers are aware that running utilities like "Startup Repair" involve certain elements of high-risk, which, if things go wrong could not be rectified by a non-technical user?

In view of the above concerns would it not be prudent for Acronis to state quite clearly in large print that Acronis True Image Home version 10 is NOT SUITABLE FOR NON-TECHNICAL USERS who are running their systems under Vista?

Once your engineers have resolved the issues and made the product 100% Vista compatible then you can remove this statement.

It would be worthwhile if Acronis included non-technical persons (e.g. administration employees, friends and family) in their test-program to test ATI Home and see if these non-technical users can use it without issues.

When I started this thread I made the observation that in my view ATI 10 was not Vista-ready. As a result of the excellent contributions by many of the best brains on this Forum, that statement now seems to be proved beyond doubt.

T.

 

MudCrab….Your sector readouts are curious and I’ll have to do some testing on this, but unfortunately I don’t have the time at the moment. I’ve not seen or experienced this so don’t know what is going on.

Mark….The BCD does indeed hold the partition offset details and targets the partition by this offset value. http://www.multibooters.co.uk/bootmgr.html

Foghorne……The MBR targets the partition by the partition table, which will be correct for any partition no matter if it was created by Vista or a third party partitioning tool. The correct offset will always be correctly written in the partition table. The MBR starts the PBR (Partition Boot Record) of the partition, which then starts the BCD.

 

McTavish:

It just occurred to me that you are the author of the excellent series of articles on http://www.multibooters.co.uk/. I want to thank you for writing these articles. I can't tell you how much I've learned from them. To anyone else following this thread, do yourself a favor and learn about how the PC boot process works from McTavish's articles.

 

Done. A good read and "Thank You" from me too McTavish.

Until Acronis fully puts this issue to bed via a new build, I will insert a link to this thread in the forum sticky titled <Product Links, FAQs & Useful Forum Threads>.

Regards

Menorcaman