Poll on Backup software

kado897 said:

But surely that is the case with any software. There are degrees of reliability but nothing is 100% reliable.

Click to expand...

Hence my description of "least unreliable".

As a side note, your source machine ought to be 99% reliable, and a good backup will pick up .99% of the slack > 0.01% probability of disaster

If your source machine is only 90% reliable, then a good backup will only pick up 9% of the slack > 1% probability of disaster

Huge difference!

I ran into a problem today replacing MR 8.1 Home version with a Free version, revealing a clear "bug" in the Free version, reported that to MR, and their reply is that they are not supporting the Free version anymore. So, I replied to them seeking to confirm that.

I understand them not adding any more features, but if a "bug" is found, they should be willing to investigate that and issue a patch. If they are not, then it really has been "retired".

Mark Phelps said:

I ran into a problem today replacing MR 8.1 Home version with a Free version, revealing a clear "bug" in the Free version....

Click to expand...

What was the bug?

When I've moved one of my licences to another machine I find the best way to go from Reflect Home v8.1 to Reflect Free v8.0 is first to completely uninstall v8.1. You can keep the user data, a fresh install of Reflect Free v8.0 should pick up your backup definitions.

kelper said:

The irony is, to test your backups, you need to do a restore; if it fails you may have to reinstall windows and your apps. EaseUS failed me due to a sector-copy write error but MR didn't. I never did find out why but EaseUs support was not helpful.

Click to expand...

That's what the "Verify image" checkbox is for. It performs an image restore without actually writing the data to disk, as instead it fully compares the data that's already there with the data that would otherwise be written. The only way to make this possible is by not relying on hot images. They have their merits of course, but there are limits to how reliable they can be in comparison to capturing a Windows partition with what I like to call "full fat" data verification after the Windows installation has been taken offline so that it can not alter the data until the image has been verified, robustly. So, cold imaging is a strict necessity if the goal is to make a reliable system backup of a non-transactional based system. Cutting corners─like how you and most other people suggest when they talk about creating a full system backup─simply breaks the whole definition of that. It is best described as pushing your luck. Some actually even go much further than this, by trying to push others to do the same. Not that you shouldn't be free to decide for yourself on whether you want to take the added risk or not, BUT... there is an added risk. Whether you like it or not. As the saying goes, ignorance is bliss. It wasn't me who invented this definition, BTW. I know that it sounds harsh. No matter still, there exists a reason why very large corporations all still adhere to fully standardized (transactional) systems and subsystems, still running on Enterprise Java even today. That is, much more heavily than the majority of IT professionals would have you believe. It is what it is.

But like I also said, a solution has to be practical enough to use. The bootable Rescue Media ISO file of Acronis fits that description in pretty much every way. Sure, having to shut down Windows is impractical. Just not so impractical that reliability should have to suffer in order to avoid this impracticality. (No, before anyone should ask, merely confirming that it still boots correctly and the apps still work is not the same as robust verification.) It [the bootable ISO file] also makes up for that impracticality, with features like the ability to pick a destination folder that is located on the actual Windows partition itself, the easy ability to specify files/folders to be excluded from image, a UI that is noticeably more straightforward when compared to Macrium Reflect, plus the fact that the image creation goes nice and fast (and finalizing the install of a Cumulative Update on Windows often takes longer than it, so...), etc.. As for EaseUS, in my book it just isn't even worth mentioning TBH.

How come I can verify an old backup? It is not identical to what's on my SSD. Anyway, how I use MR is not to preserve data, it's to reinstall windows and apps quickly and that is how many people use it. YMMV.

kelper said:

How come I can verify an old backup? It is not identical to what's on my SSD. Anyway, how I use MR is not to preserve data, it's to reinstall windows and apps quickly and that is how many people use it. YMMV.

Click to expand...

Verification is done by comparing a hash of the data in the image with a hash created and stored at the time of the backup on a block by block basis. So it will fail verification if either the hash or the data becomes corrupted.

That can't be correct. I reckon verify just checks the image for integrity. But I'm guessing.

kelper said:

That can't be correct. I reckon verify just checks the image for integrity. But I'm guessing.

Click to expand...

You guess is wrong, and @kado897's description is correct.

Backup verification checks the entire contents of backup files against MD5 message digests (Hashes) created from the source data when the backup was created. If data is read back with a different MD5 hash to the hash that was generated when the data was written, it is guaranteed to be corrupt.

Click to expand...

Hdmi's description is wrong, kado's is correct and confirms what I said.

kado897 said:

Verification is done by comparing a hash of the data in the image with a hash created and stored at the time of the backup on a block by block basis. So it will fail verification if either the hash or the data becomes corrupted.

Click to expand...

When the Verify image checkbox is checked during image creation in the bootable ISO of Acronis (at the bottom of the progress window that appears the moment when the image creation task is started), it performs a true comparison of the actual data itself. This was a design choice that they made to not only verify the data, but to verify the integrity of the storage hardware also in addition to this, i.e. by reading all the data a second time. The chance of a read error going twice undetected yielding the same data twice in a row is pretty slim. There is no other way to check this checkbox, as this checkbox simply doesn't appear anywhere else. Once you boot back into Windows, the source data will be altered so then it can no longer fully be compared with the data captured.

kelper said:

Hdmi's description is wrong, kado's is correct and confirms what I said.

Click to expand...

I am deeply sorry, but I read through remarks like this as if claiming that there are no colors in a rainbow.

hdmi said:

I am deeply sorry, but I read through remarks like this as if claiming that there are no colors in a rainbow.

Click to expand...

There are only two colours for a severe protanope.

Personally I use Acronis as it came with my 1TB External Hard Drive. It has not failed me yet.

JesseRBassett said:

Personally I use Acronis as it came with my 1TB External Hard Drive. It has not failed me yet.

Click to expand...

Been using Acronis since 2009, it's been great, so I have no interest in changing something that's been working.

RD Mercer said:

Been using Acronis since 2009, it's been great, so I have no interest in changing something that's been working.

Click to expand...

@RD Mercer ,
Have you found Acronis is a little slow lately? For example I am cloning my ssd to my external hdd and it says it will take 4hr and 29 minutes.

What is the total GB to back up? Sine qua non.

Has anyone tried O&O DiskImage 18: Backup Software for Windows?
They do have very good software (not free)
Their O&O DiskImage 18 cost US$ 50 that is less than MR that cost US$ 70

O&O is for one PC, the MR license is for four.

kelper said:

O&O is for one PC, the MR license is for four.

Click to expand...

4 licenses cost $140