Verifications in Macrium Reflect Free (MRF)

I'd appreciate comments from the forum experts on the following, thanks!

The most real verification is doing an actual restore of the MRF image. You can do spot checks on OS, apps, user data. But if the image is bad, you may have permanently damaged your user data, for example (hope you have a good image at hand, it does not have to be an MRF image, it could be an image from another backup software)

One step down the MRF verification list is viBoot. You can do spot checks on OS, apps, user data. And if the image is bad, just delete it (and the VM) and try again.

Two steps down the MRF verification list is mounting the image in File Explorer. You can do spot checks on your user data. Protecting user data is most important for me personally (protecting OS and apps is less important for me because they can be reconstructed) and having said that, why use such a feature-rich app like MRF to do the user data job? Because MR is very reliable? I note that MRF does not have an error log for user data.

Three steps down the MRF verification list is to push the verification button with the 'automatically attempt repair' option.

Bottom of the list is not doing any verification of the MRF image at all. I guess that this may already give some people their peace of mind while others will want one or more verifications done at least before and/or after making important changes.

You have not included the optional write verification that is done at time of creation of image.

Also, restoring rhe image in a native booring vhdx file is the best test short of a full restore, as that 100% uses native host drivers rather than interface vm drivers.

I seriously doubt most users go beyond doing the write verification and perhaps using ViBoot.

After that, I am sure most users basically take an "experience" leap of faith that the restore will work fine.

In the end, data integrity is more important really than image backups of OS and software which can be rebuilt.
Lost data can be irretrievable of course.

The first two golden rules of computing are
1. Always back up data regularly.
2. If there is any doubt, refer to rule 1.

Conventional wisdom is critical data should preferably be backed up on two (or more) different media, preferably at different locations. My personal choice is a usb hard drive and OneDrive. Others will have different choices.

Less critical data should be backed up on at least one other media e.g. usb hard drive.

Non critical data - well that is for each user to decide e.g. if it is videos that can be redownloaded from web, is it worth backing up?

I am a great Macrium Reflect fan and I rely on it 100% for OS + apps backups BUT I do not rely on it 100% for data (integrity) backup. I just use it as part of my overall data backup strategy.

There really is no "one size fits all" strategy to backups and verification. Every user has to decides what suits them.

As an aside, all you say above is equally applicable to any of the mainstream imaging tools and virtual machine tools. Even though they may not have viboot, you can still basically do same i.e. restore an image to a vm but it is a (less slick) manual operation.

cereberus said:

You have not included the optional write verification that is done at time of creation of image.

Click to expand...

I think I included that as the following.

Haydon said:

Three steps down the MRF verification list is to push the verification button with the 'automatically attempt repair' option.

Click to expand...

-----------------------------

cereberus said:

Lost data can be irretrievable of course.

Click to expand...

That's why I regretted that MRF has no error log for user data.
-----------------------------
I think it is time to go beyond backup basics, from 'doing a backup' (that has been discussed ad nauseum) to 'doing a backup plan' (that has been barely discussed) like using MRF without error log for user data with robocopy with its error log for user data.

I mentioned in scattered threads using multiple backup tools (MRF, robocopy, FH, Direct Copy are the ones I use, there are obviously many, many others) and using multiple backup media (multiple DAS is what I use, there is obviously NAS and cloud) and combine tools and media into an effective and still efficient backup plan. Perhaps we can continue this thread in this 'backup plan' direction, it's an extremely broad topic, though.

@cereberus I use the paid version of Macrium which lets me set my default option to verify each image as it is is created. Does the free version not have that automatic option?

Screenshot from Macrium Reflect 8 Free, clicking on the Verify button ...


... displays the following window where you could tick off 'automatically attempt repair'


I don't see a setting where I could have included the verification step as part of the image creation step. Anyway, that's what I meant in my OP.

I see. So with the free version you have to manually run the verify. Here's my version.

Haydon said:

Screenshot from Macrium Reflect 8 Free, clicking on the Verify button ...
View attachment 29629

... displays the following window where you could tick off 'automatically attempt repair'
View attachment 29630

I don't see a setting where I could have included the verification step as part of the image creation step. Anyway, that's what I meant in my OP.

Click to expand...

Right after the backup completes is when I do the verification. Free and paid.
I think the only things I ever auto-verify are CDs or DVDs when I burn them.

Ghot said:

Right after the backup completes is when I do the verification. Free and paid.

Click to expand...

What @glasskuter meant was not 'after' but as an integral part of the image creation step.

@glasskuter I see the settings window now too, see my screenshot below, which is equivalent to yours. However, this settings window does not have a tick box 'automatically attempt repair', it ought to be still there, somewhere

Haydon said:

However, this settings window does not have a tick box 'automatically attempt repair', it ought to be still there, somewhere

Click to expand...

Update: I turned on 'Auto Verify Image' in the settings window > then created such an auto verified image.

The 'Verify' button is still available for the auto verified image, and with that the 'automatically attempt repair' is still available in case there is an issue with the auto verification. That seems to be the design intent.

I wonder, though, whether it would not be much more prudent to delete the problematic image altogether and simply create a new image that passes auto verification right away.

Haydon said:

What @glasskuter meant was not 'after' but as an integral part of the image creation step.

@glasskuter I see the settings window now too, see my screenshot below, which is equivalent to yours. However, this settings window does not have a tick box 'automatically attempt repair', it ought to be still there, somewhere
View attachment 29633

Click to expand...

I have that enabled on the free version and so far it completes ok. I haven't yet come across what options or actions it would show if verification failed.

Fabler2 said:

I haven't yet come across what options or actions it would show if verification failed.

Click to expand...

Well, I did very mean things LOL in testing MRF, like deliberately disconnecting the external backup medium for a few seconds to induce a failure in creating or verifying an image. All options or actions are still there, MRF is VERY robust.

Edit: Don't do those things on your main work computer. I did that on my test computer.

One thing you may want to consider, especially with critical data backup, is to keep at least one uncompressed copy of the data.
Although most compression tools don't tend to go the way of the dodo, often there is a chance that the app you use may have a bad update causing your backups to become unusable for some time, or a bad system update causes issues.
This can be inconvenient to disastrous depending on the type of data backup and the use case and whether you have kept an older version of a start up disk.
Any manipulation of data from just copying to a new location to compressing then copying/ moving runs the risk of data corruption, it is also a lot easier to recover data from an uncompressed state.
Another valid backup method for critical data, depending on the type of course, is to have a hard copy. A folder of A4 printouts of login info, passwords, favourite website addresses, config info etc is (barring theft, fire and acts of mythical divine beings) a lot more resistant to corruption and failure than any computer based technology.
These methods are much easier to verify by their very nature as well, although can be a lot more work. :)

Here's another - but longer indirect way to test and you can run the whole thing from an external drive - and it's a physical test so all drivers etc and data can be checked for integrity. This way you don't corrupt your "C" drive if the backup image is faulty and works on Windows HOME as well as other editions of Windows so doesn't need any HYPER-V settings or VM settings in Windows. Steps 1) to to 5) only need to be done once to "prep the environment".

1) attach external drive to computer
2) in command mode go to diskpart
3) create on the drive 1 EFI partition size 100 formatted fat32, 1 msr partition size 128 rest of drive one NTFS partition.
4) create a vhdx file (virtual hard drive file of the size of your current windows drive -- it can be smaller if you want to "shrink" the size and there's enough space) --
5) select and attach the vhdx. select the disk, format it as a SINGLE ntfs partition (no efi etc)
exit disk part.

Now run your restore to the newly assigned disk - you only want the basic Windows partition - not any efi or recovery partition.
then we need to install the bootloader so
Boot stand alone macrium recovery or a windows iso and get into command mode
assign the 100 mb "hidden" partition on the external disk as letter S
attach the vdisk as letter W (this is the one with the windows image on it).

then cd w:\windows\system32
bcdboot w:\windows /s S: /f UEFI

exit everything and boot the external drive -- now you've got a backup and running as a "Windows 2 go" system. !! Test to your hearts content.


looks complex to do but once you've done it it's surprisingly easy. Also once ve got a spare useable physical Windows system for testing software / hardware without needing a VM or breaking anything on a running machine. If the external device is large enough too you can have several copies of Windows on it too - all different versions if necessary and they'll all be activated with digital license.

Cloning usually works easier as it's easier dragging / dropping a partition and re-sizing so on restore you just clone the external Windows image (not the entire disk but the "disk windows is on") , run the macrium fix boot problems and boot your orginal C drive system again.

But however you do it one should always attempt a restore once in a while -- just because a backup says completed you can't be sure that a restore will be 100% successful. It usually is but data errors, backup media corruption or whatever can cause these things to fail and absolutely 100% according to that law that's more rigid than Einsteins E=MC squared equation the restore will fail on the one time you absolutely must have it.

Cheers
jimbo

glasskuter said:

@cereberus I use the paid version of Macrium which lets me set my default option to verify each image as it is is created. Does the free version not have that automatic option?

Click to expand...

same

@DigitalGoat Those are exactly the reasons I still use the 'primitive' Direct Copy for 'irreplaceables' (only a few GBs) on USB and DVD 'archives' in addition to regular backups. I check the 'archives' only every other year or so, just to see if they are still good (Brownian motion in storage media ) but they are safe from malware, safe from frequent use :):) ... and then there is still room for paper in a safety deposit box

@jimbo45 Very true comments about doing actual restores once in a while, that's part of the reason for the hierarchy in the OP, from doing actual restore step-by-step (in 4 steps) down to not doing any verification of the image at all.

Haydon said:

@DigitalGoat Those are exactly the reasons I still use the 'primitive' Direct Copy for 'irreplaceables' (only a few GBs) on USB and DVD 'archives' in addition to regular backups. I check the 'archives' only every other year or so, just to see if they are still good (Brownian motion in storage media ) but they are safe from malware, safe from frequent use :):) ... and then there is still room for paper in a safety deposit box

Click to expand...

I think one has to draw a distinction between backup and archiving. Backups are really a (mostly temporary) activity to enable a user to revert if something craps out. I do not hold image backups going back more than 2 or 3 builds, as I will never use them again. Data is different (some I want to keep for long term, some I probably keep for a while, and some are disposable),

Archiving is a completely different kettle of fish e.g. solicitors have to archive legal agreements, practically in perpetuity. This data would typically be stored in some form of database, often managed (for a fee) by a 3rd party. One of MS's major revenue streams comes from Azure archiving for example.

Of course, life is not that black and white e.g. backing up treasured personal photos to onedrive - is that backup or archiving?

I would say it is more backup (for long term), as it is usually just a copy online, but not incorporated into a specific database, which is fine for relatively small quantities of data, but after a while, it is like looking for a piece of hay in a needle stack (painful).

I read some more about Hyper-V and the fact that Windows no longer runs on bare metal became more and more disagreeable to me > I turned off Hyper-V > I am now running viBoot VM on VirtualBox

viBoot VM on VirtualBox is as convincing as viBoot VM on Hyper-V in terms of verifying an MRF image :) just short of doing an actual restore.

When I opened an MS Word file, a message from MS 365 reads something like 'It looks like you have new hardware, please connect to the Internet so we can verify your subscription' I am surprised that MS Word files notice the difference WOW Lucky me, I did not turned on the Virtual Switch when defining the viBoot VM LOL

No other suprises so far in the very short time since I am running viBoot VM on VirtualBox, but stay tuned

Haydon said:

I read some more about Hyper-V and the fact that Windows no longer runs on bare metal became more and more disagreeable to me > I turned off Hyper-V > I am now running viBoot VM on VirtualBox

viBoot VM on VirtualBox is as convincing as viBoot VM on Hyper-V in terms of verifying an MRF image :) just short of doing an actual restore.

When I opened an MS Word file, a message from MS 365 reads something like 'It looks like you have new hardware, please connect to the Internet so we can verify your subscription' I am surprised that MS Word files notice the difference WOW Lucky me, I did not turned on the Virtual Switch when defining the viBoot VM LOL

No other suprises so far in the very short time since I am running viBoot VM on VirtualBox, but stay tuned

Click to expand...

I rather think you are overstating your case - having the hypervisor on has no really significant real world performance impact, is properly segregated from other vms in an operating viewpoint, and running Windows vms in Hyper-V are definitely more efficient than running in Virtualbox (that is not just opinion - I have done tests).

With Hyper-V, Windows still effectively runs on bare metal (i.e. uses exactly same drivers) but so do its vms to some extent as well.

With virtualbox, vms are totally isolated from the hardware using emulated drivers but there is a penalty i.e. virtualbox itself uses much more resources.

Not only that, there have been a number of cases where virtualbox has crashed following Windows updates. I have never seen that happen with Hyper-V for years.

The only time when virtualbox is preferable (in my opinion) is when running OSs that cannot be run in enhanced mode with Hyper-V.

I have just run passmark benchtesting with hypervisor on and off just to check on my new laptop, and reconfirm above,

The average figures are virtually identical.

It is hard to interpret results precisely, as figures get weighted down by crappy graphics.

It seems the CPU is slightly reduced with hypervisor on taking a bit of the CPU performance, but in the end, that does not mean much as I am rarely running my I7 CPU anywhere near flat out.

Same is true to the drives - nvme drives but again, I only run them flat out when copying large files (many GB).

In other words - real world impact to me is barely detectable unless doing some really intensive memory/disk work e.g. intensive video editing.
Gamers might find the impact more significant I guess.

I accept the fact that running with hypervisor off maximises theoretical performance if not running vms, and you only get a hit when actually running virtualbox.

However, if you run vms a lot as I do, the real world minor performance loss of having hypervisor on is far more offset by vms running in Hyper-V as opposed to Virtualbox.

Of course, it is each to their own.


Hypervisor On


Hypervisor off