Macrium Reflect

Bug Fixes v7.2.3825 - 2nd November 2018

http://updates.macrium.com/reflect/v7/v7.2.3825/details7.2.3825.htm

 

Get a 'Could not start the setup' Launcher Error pop-up in the Free version, maybe because the fix does not apply to Free ...

 

I went back to 7.1 and thought that I will wait for a while before upgrading, so I didn't try upgrading it yet.

 

Reflect Free does not include CBT, so it would be unaffected.

 

You don't HAVE to create new Rescue Media since there haven't been any fixes relevant to the Rescue environment mentioned in the release notes. Normally those are called out in red text when they exist. But you shouldn't expect a popup dialog either. The new Rescue Media Builder works a bit differently compared to the old wizard. If you open it and check the Status readout at the top, it will tell you that the Rescue Media needs to be rebuilt because there's a new Reflect version available, and possibly a new WinPE 10 version if you're using that (7.2 uses WinPE 10 1709 rather than 1607). But it won't prompt you to do it. In fact even the old Create Rescue Media wizard didn't prompt you to rebuild unless you stepped through to the very last page.

 

At least up until now, version 7.2, I judged installing Image Guardian to be riskier than not installing Image Guardian - especially since I image to two redundant external hard drives that are online only for imaging. I'm also using Bitdefender on all five of my machines. Image Guardian and malware are both evolving. I'd be interested in the views of others on whether Image Guardian is better installed or not installed.

 

A beta period is not a substitute for actual quality control. Here's hoping you never need medical care that uses a device whose developers think as you do. "Let's get the kinks out by harming a few patients. What else can we do?"

Many of the problems with this release appear to be packaging issues that would have been avoided by releasing as an uninstall/install sequence rather than an update. That's just rushing software out the door without due care as to how it's shipped. The cost to Macrium of being more careful in packaging would be a great deal lower than the cumulative cost of the entire customer base burning labor hours on workarounds.

 

Small business environment. W7 (32), W10 Home 1803 (64), W10 Pro 1803 (64). All have been reverted to 7.1 except one of the Pro machines. Webroot SecureAnywhere.

It seems that many of the issues are related to MIG and CBT, and possibly complications having to do with their pre-update status. This is the sort of thing beta testers are unlikely to encounter unless they are installing the beta software in the actual scenario the end-users will be having (new point release over old).

 

I personally use Image Guardian and jettisoned third-party anti-virus years ago. I gave up third-party anti-virus years ago because Microsoft's built-in Defender has gotten quite good, and everything else on the market seemed to be some combination of bloated, expensive, "noisy" in terms of notifications, or problematic in terms of interfering with legitimate activity. There are several cases just on the Macrium forums of anti-virus doing things like blocking or dramatically slowing down image backup creation, causing Rescue Media creation to fail, etc. That's probably why Macrium has this KB article. And that's all before considering the multiple cases where third-party anti-virus has turned out to be the cause of Windows updates or new Win10 release upgrades failing -- because the AV solution had hooked the OS in some way that Microsoft didn't recommend or support, so when MS changed something that nobody else should have been using, the AV solution broke and took the whole system down with it. Then there are multiple where anti-virus has become a system vulnerability it had a bug in its code, and since anti-virus by its nature constantly scans files and incoming network traffic, those types of bugs are exploitable in a wide range of ways. And most of these anti-virus vendors expect you to pay for these "privileges". No thanks.

I don't think TheRollbackFrog suggested that a beta was a substitute for quality control, but the analogy you described is EXACTLY how the medical industry works. That's what clinical trials are. Of course the first rounds don't use humans, but there is always a human trial period that almost always has harmful side effects on a few patients -- because that's how things get discovered. And then even after the drug gets approved, there are inevitably more harmful side effects discovered later thanks to the much larger sample size of the general public.

It's the easiest thing in the world to say, "This or that should have been caught" when you have the benefit of having the problem present itself to you as opposed to having to look for problems proactively, never mind considering all of the things that were already fixed in beta. It's impossible to test all code paths and all scenarios, and it's not as if everyone who did a "new point release over old" upgrade from 7.1 to 7.2 is having the issue that you are. It's not clear what caused your issues, which means it's unclear how common a scenario it is and therefore how reasonable it is to expect that Macrium should have caught it.

And frankly, if the worst issues with 7.2 are a) some problematic update scenarios that can be worked around by a simple and painless uninstall and reinstall, and b) a CBT issue that created long shutdowns and that was just fixed less than 48 hours after the initial release, then I consider that pretty successful. But obviously you don't, and that's fine. But if that's enough to get on your "This should never have been released!" soapbox, then you should probably refrain from upgrading to the initial builds of lots of point releases given the realities of software development and release strategies these days. See Windows 10.

 

We've been running Image Guardian, but may not continue after updating to 7.2. We keep copies of image backups offline and offsite where they are pretty secure. It does protect the interim file and folder backups, however. As for CBT, I haven't seen a particularly noticable performance benefit to it. It seems mostly to be gold plating.

 

CBT eliminates the "Looking for changes" step of Differential and Incremental image creation. On systems that experience high data churn and/or go significant periods of time between backups, that step of cataloguing all of the changed blocks on the partition since the last backup can add quite a bit of time to total image creation time -- and therefore having CBT keeping that journal current in the background can therefore save quite a bit of time. But if you disable CBT and find that the initial "Looking for changes" phase is consistently short enough for your purposes in your use case, then don't use CBT.

 

Your endorsement of Image Guardian is a confidence builder, and I think I'll probably end up installing it this time. Your good opinion of Defender is something I didn't hear often when I last researched anti-malware software a couple of years ago. It's worth another look. Thank you for the good information.

 

@AlphaOne - I have never used Image Guardian as I have a 3rd party driver which performs basically the same exact thing but for a much broader use scenario. It was in use long before MIG came along.

As a very frequent Incremental image taker, I have found CBT to provide exceptionally small and speedy Incremental images as well as faster restores when using those CBT created images. Since I use 15-min workday intervals on my Incremental requirements and I take many MANUAL "snapshots" during the day when the System is in a testing scenario, CBT has proved very useful for me. If your Reflect storage mgmt is designed to do no more than single daily Incrementals, CBT does not offer any significant advantage for its user.

If Macrium had ever made available its Continuous Incrementals feature it boasted about in February 2017 as part of its v7 release, I would have been a very happy camper. But alas, it never came to fruition and totally without explanation from Macrium. I can always hope, though.

 

Another appreciated vote of confidence in Image Guardian. The difference between what you and I are doing, if I understood you correctly, is that once you've created an image it gets permanently taken off line. I keep all my images on those two external hard drives (except for infrequent optical storage) which I repeatedly put back online and therefore plausibly in harm's way.

 

To my knowledge (and based on Macrium's own descriptions), CBT doesn't affect the data that makes it into the image, so I don't see how it would make the resulting images smaller or affect restore times. All it does is keep a continuous journal of the blocks that have changed on the system so Reflect has that information available upfront rather than having to figure it out when the user requests an image. If CBT affected the end result of the image, then that would be cause for concern, since by definition two different results couldn't both be accurate representations of the source. The only time I know where CBT affected the end result was during the 3 builds where a bug caused CBT to generate corrupt images when it was used, and that quite obviously was cause for concern.

 

Well, it happens that my consulting firm specializes in process improvement for software development and services. Our clients tend to be involved in critical software such as defense systems, the space program, and medical devices. I'm not sure where you came by your authoritative viewpoint, but I can assure you that you are mistaken. Considerable rigor goes into developing software for even (comparatively) simple devices like blood glucose monitors.

I will admit that back in the 1960's (yes, I was working that long ago) during development of the earliest heart monitors, prisoners volunteered to be test subjects, In exchange for a day spent out of jail, hooked up to a monitor, they were paid a carton of cigarettes. But, in those days, most everyone at the heart institute smoked like a fiend.

 

I stand corrected . I just went back over my notes taken during the BETA CBT development process under v6 and what you state is indeed correct. I guess it just "seems" to generate smaller Incrementals. The Incremental speed, though, is definitely much faster in my storage mgmt scenario.

Thanks for keeping me in line...

 

You opened this discussion with, "This software should never have been released!!" and you're calling out others for taking an authoritative viewpoint? Well ok then. As it happens, my wife has been running and auditing clinical trials around the country for quite some time, so I have a decent window into how that side of the medical industry works from a testing perspective -- although frankly my case could be supported purely on the basis of empirical evidence from reading news reports, where examples abound of negative side effects of drugs manifesting in human trials despite not manifesting in earlier animal trials, and then others manifesting after approval that never showed up even in the human trials. Your case boils down to, "The 7.2 initial public release had some issues, some of which affected me [though none of which were major or impossible to work around]. I think they should have been caught before release. Therefore I believe Macrium doesn't employ rigorous software testing." Ok.

And even that entirely ignores the obvious fact that the stakes are rather higher in the defense, space, and medical industries you mention, so one would EXPECT testing to be more rigorous there. Nobody was put at risk of dying because they had to uninstall and reinstall Reflect rather than update, or choose between not using CBT and having to wait a few extra minutes for their system to reboot for the 48 hours between 7.2's initial launch and the time that issue was fixed. And again, it's not even clear how common the problem cases you're talking about are, so the "egregiousness" of those bugs making it into production can't reasonably be judged. Maybe if Macrium could charge an amount for Reflect that would give it the profit margins realized by the businesses that serve those other industries you mentioned, they'd have testing up to your standards.

Anyhow, I'm done. This is wholly off-topic at this point and therefore unfair to everyone else watching this thread. If you're unhappy with the quality of Macrium's releases, then find some other solution that meets your needs and whose release quality you consistently deem up to snuff.

 

I like that it's optional. I find that in general, the more functions a given tool or piece of software is expected to perform, the less effectively it can perform any one. However, I admit that sometimes I hate having to find, learn, and keep updating a different app for every little thing that I want to do, so occasionally I appreciate getting additional features in my existing apps that increase their versatility. I personally wasn't aware of Pumpernickel, but I'll keep it in mind. But in the meantime, although I haven't had a ransomware attack on my PC to really test MIG, it's not creating any problems for me either and it has blocked everything I've done to test it, so unless I see reports of MIG failing to block a ransomware attack that Pumpernickel stopped, I don't feel any need to switch. But many people DON'T go searching for applications every time they think of something that might be useful, and some people don't even realize they want something until they see it offered to them. For those people, I think having Image Guardian as an optional part of the backup solution they were already using is a value-add, so I certainly don't fault Macrium for committing the engineering resources to develop it.

Also, can Pumpernickel block or allow modifications to data being shared out on the network based on the process on the remote PC that's attempting the modification? MIG can do that, i.e. if I have a Windows file server hosting Reflect images where MIG is enabled, then my file server's MIG will allow Reflect on other PCs to modify the files it's hosting, but block other processes on that remote PC from doing the same thing. I don't know how it does, because to my knowledge a file server can't normally tell what process on a remote PC is being used to access the files it's hosting -- and when I asked, Macrium said that was proprietary -- but somehow it works.

 

Thank you Froggie and Peter2150. Being your preference is a strong reason to think it might be a better alternative for me too. I'm going to take a good look at it.

 

Thanks jphughan. When I upgrade to v7.2 is it better to use the internal updater (the one that pops up when Reflect starts) or to uninstall and reinstall? I use the free version of Reflect so I don’t have to worry about the license key.

 

I just finished reading the Pumpernickel FIDES thread. It seems that at minimum FIDES might be the most effective and efficient way to protect my external drives that I use exclusively for MR imaging and FreeFileSync file backups.

 

Thanks Pete!