Reliability of "live" imaging

I didn't notice that anyone said otherwise. Recursive file-by-file comparisons of an image against its restoration could prove if they were the same (or not) much more persuasively than huffy assurances grounded in supposition rather than actual knowledge of the product internals. The fact remains that I have encountered puzzling anomalies following restorations. I'm delighted to know that you have not.

 

I do use VM software, such a VirtualPC 2004 and 2007, VirtualServer 2007, VMWare Workstation and Server 5, and VirtualBox, and a few others too.

But some things just cannot be done in a VM, period, and then some things can't be done in a VM with reliable and repeatable results because the host OS being busy at random times, etc.

Hopefully that answers that question. I had an E: drive, separate 120GB Seagate SATA drive, that has 47 different images on it for those VM applications, all used regularly for various testing purposes.

My computer doesn't sit idle for long at all, I can assure you.

bb

 

Backing up whil running in Windows is at least as safe as rebooting after a crash.

When you do a backup, the imafges are a snapshot as of a point in time, so you can restore to that point in time. No different than if the system crashed, or whatever, andyou had to reboot, in which cases the files would be in an unknown state from a particular point in the time.

In an extreme case, you might not be able to reboot.

If you backup from outside of windows, then no files are running while you backup, but it is still only a backup in a point in time.

I would expect this to make no difference on other than, say, a transaction processing system that just cannot afford to get particular files out of sync. Although, the better applications likely protect themselves.

Indeed, you can even boot to Linux and use a Linux backup program.

Do whatever makes you feel better.

The SEagate product is DiskWizard, and it can only be used if you have at least 1 Maxtor or Seagate drive.

I tried Disk Wizard recently,it is a watered down version of TI 10, but it omits the ability to define jobs that can be manually/automatically run at a later time. I use 4 jobs, one for each of my backup drives. It would be too tedious to manually create a new job each time. Do not recall what else is different.

 

For such comparisons, you can use:

GetFileTypeDistribution
CompareDrives

 

There is no problem comparing data drives but how do you compare an OS drive with an image ?

If I make a "live" image of C: how can I compare C: with the image ?
As the process of shutting down the machine changes C: and it is worse still if I boot to C: to compare with the image. Using File compare programs I have been able to see that both the image and C: are "essentially" the same and for all practical purposes are the same. Is there a program that could be run from another machine that could compare C: and the image without changing C: ?

 

Actually, I have used your CompareDrives program. Didn't know about the other one. Unfortunately, such tools aren't helpful in accounting for the registry anomalies that have left me scratching my head.

 

See the description of the following:

GetFileTypeDistribution
CompareDrives

 

Peter I understand what you are saying but:

Having made the system image what happens to C: when you mount the image ?
If C: is involved it must change even if only by a small amount

 

Hello layman,

Can you just confirm that you have been definitely talking about registry anomalies after a System Disk/Partition restore rather than a Files and Folder restore to your system partition?

Regards

Menorcaman

 

I not worried at all. I accept that Acronis makes a valid image every time - It's just difficult to prove. I was trying to think of a way to show layman that Acronis makes perfect images. Using Beyond Compare 2 http://www.scootersoftware.com/ I have just made a system image, then mounted and compared C: with the image. 30,562 files out of 30,612 were reported as being exactly the same. The remaining files were ones you would expect to change. The act of comparing will itself change files.

 

I imagine the TI log file changed

 

While you are booted toWindows, files will be a changing.
The only thing you can do is use programs such as the following:

GetFileTypeDistribution
CompareDrives

And see what differences are noted.
Most can be easily explained.

The longer you wait after a backup, the more changes you will see.

 

Yes (confirm that.) I evaluated the files/folders feature when it first came out, but found it lacking. I use ZipBackup for that purpose.

 

I have done the same sort of test a number of times, but you have to recognize that it proves only what it proves. A more thorough test requires two stages. You not only have to prove the integrity of the backup image, but the integrity of the restoration process, also. I've done this second stage test more times than I've done the first - without ever pinpointing an inexplicable difference. Unfortunately, from a software quality control standpoint, you need to run both stages on the same test case. And even doing that might fail to recreate the conditions that are the root cause of the issue.

I've got no explanation for the wierdness I've encountered more than once, but nor could I think of any way operator error was to blame. Can't happen? I'm glad you find that explanation reassuring.

 

I'm replying to my own message because I've once again encountered the sort of anomaly I experience with TI from time to time. Yesterday, I visited the Windows Update site which indicated that I needed to download a revision to the update facility. Since I had fresh images for the two machines I'm running, I proceeded to download the update on both. On one machine, the update didn't appear to have installed properly (I got an error in Windows Update), so I restored the backup image and re-ran the WU update successfully.

Now, here's the thing: Several of my on-line banking accounts use a special form of cookie for security verification purposes. When I next attempted to use these accounts, the security verification failed. Looking at the cookie directories, there is nothing obvious amiss. Obviously, the banks don't publish the specifics of the security technique they use, so I don't really know why the verification fails.

I mention this incident merely to caution TI users against the assumption that a restored image PRECISELY recreates the state of the machine at the time the image is cut (and I mean the logical state of the machine, not the physical representation.) The differences are subtle, but I've experienced enough of this sort of thing to conclude that Acronis takes a few short-cuts in the backup process. It hasn't yet caused me enough trouble to consider that it offsets TI's good qualities, but it certainly makes me uneasy.

 

It's not really Acronis, when you restore an image, some web sites detect the date anomalies in the stored cookies and expire them. This can occur with various sites not just banks. In simpler terms, the cookies contain information that is too old as compared to what the site has recorded about your visits.

 

I agree, IMO, it just means TI actually restored an old file. I have never had any reason to believe my data contents have been strangely modified.

I don't know if my bank still does it but years ago when I was having trouble logging in they made me correct my computer's time which was out by about 5 minutes - don't think it fixed the problem but it was in the equation somewhere.

 

That can't be the explanation, because the backup image was only an hour or two old.

 

If you cannot accept that the Acronis imaging technology works then of course feel free to add on your own or third party checks and balances to prove whether the program works or not for you.

I frankly think you are looking for "Ghosts" that do not exist. I have had, in thousands of successful backups, the odd reported backup failure. In my ignorance I assume that Acronis have built their own controls into the imaging process to ensure accuracy and completeness. I make this assumption with some conviction because otherwise what would cause the program to report a failure?

I changed my backup hardware over a year ago and since then I have restored between 400 and 500 complete images. Never once have I found or suspected that the restored drive was not identical to the source drive. In that time I have had 3 or 4 restores fail. I have not bothered to find out why as I only restore to another drive leaving the source drive untouched until it is due to be recycled.

Xpilot

 

That is contrary to the nature of this particular security mechanism. Very likely, the bank's security technique makes provision against tampering, but that is precisely my point. Something in the nature of restoration appears to trip this safeguard. In other words, the mechanism recognizes that there is a difference between the state of the machine at the time the image was cut and the restoration just an hour or so later.

 

Unfortunately, I've never experienced any of these peculiarities at a moment when I could devote time to carefully researching and documenting the incident. I've continued to use TI because it's fast and the design choice of a virtual disk format for images is elegant. It serves my purpose well enough, most of the time. I don't entirely trust the product, though. It is rough around the edges and has an extremely amateurish user interface design. The latter is a biggie mainly in that it makes you wonder about the quality of the product's internals. Speaking as someone with four decades in the field of software as product, I think Acronis has a lot to learn about usability and software quality control. I switched to TI when Symantec bought PowerQuest and (inevitably) trashed Drive Image. I remain wary of it, however, because of the occasional anomalies. Kneejerk defenses by product loyalists don't impress or reassure me one bit.

 

Hello layman,

As this is the official Acronis True Image support forum, please detail precisely what it is that you require support for (hypothesizing and generalizing doesn't count I'm afraid)?

Regards

Menorcaman