Acronis True Image costs performance when not used

Acronis True Image costs performance when not used(randomascii.wordpress.com)

147 points by juanviera23 271 days ago | 36 comments

userbinator 267 days ago |

but they shared symbols for the tishell64 DLL to help us understand what was going on.

This is perhaps the most surprising sentence in this article. A proprietary software company sharing debugging symbols (which are basically the closest thing to source code that isn't source code) when you just ask their support about an issue is something that I'd never expect to happen.

brucedawson 267 days ago | |

I was surprised also when this happened, but pleased. The trust was nice. The fact that they needed to give me the symbols in order to understand the issue was surprising. Why couldn't they find the calls in their source? Or analyze the trace themselves? Curious.

eps 266 days ago | |

Symbols for Windows system DLLs has been available for ages, downloadable in one click directly from within Visual Studio. So it's not without a precendent.

kg 267 days ago |

Windows ends up being a thousand cuts situation when you have enough software like this installed. If you want a responsive and stable system you end up spending a good amount of time troubleshooting freezes or mystery CPU/disk usage.

As of yesterday (this started suddenly) any time Backblaze is performing a backup my whole system gets bogged down for no reason. Was fine days ago, so who knows what automatic update for what app caused it. It's probably an interaction with some other filesystem filter on my system, like Defender or Acronis True Image (which I am stuck using since every other disk imaging tool I've used is worse). Seeing this post on the front page has me wondering whether it's actually Backblaze's fault or it's Acronis, now...

I've previously tracked down random system freezes caused by some part of the NVIDIA driver interacting with one of my monitors, where the first unlucky process on my system to call certain graphics APIs after a boot or wake-from-sleep would cause the video driver to hold a lock and enumerate display modes for 10+ seconds. The end result was a "frozen" PC still responding to input, just unable to send new pixels to any of its displays. The fun part of that one was that profiling would blame these freezes on whatever user mode process was unlucky enough to make that graphics API call - Steam's browser overlay, Visual Studio's WPF rendering, the list goes on. The troubleshooting process for this one was expensive because I got to the point of replacing components before I figured it out.

I'm curious whether people who daily drive mac or linux encounter these sorts of system configuration gremlins. My experience using linux on servers and VMs has had some situations like this but I don't know whether the everyday experience just running Linux or OS X on bare metal is actually perfect at this point.

perching_aix 267 days ago | |

> I'm curious whether people who daily drive mac or linux encounter these sorts of system configuration gremlins

Friend of mine dailies Arch. He described a few days ago a situation where he stopped being able to take screenshots from one day to the other, eventually finding out the root cause thanks to this fix: https://github.com/flameshot-org/flameshot/pull/4127

Whenever Linux package management and shared dependencies are mentioned, people always seem to be quick to mention how it's good for security that you can upgrade dependencies without the consent and knowledge of the developer of each software you're using. To me, issues like this come to mind instead. Or how I heard multiple times acquainted developers complain that packagers modified their software downstream and broke it in some ways, only for users to come complain to them instead of to the distribution.

There's a recent(?) trend of circumventing distro repos and static linking things though, and I work with Go software pretty often, which I believe static links by default: they're a joy to deploy and use. Just nice, single, self-contained binaries. But it's really not the "one true Linux way" from what I can tell, so I'd imagine it's pretty against the grain, which should tell you a lot about the situation overall. I know I'm not itching to switch, and these are only a sliver of my gripes too.

ninkendo 267 days ago | |

> I'm curious whether people who daily drive mac or linux encounter these sorts of system configuration gremlins.

IMO these issues occur any time you have third party software that does the job the operating system should be doing itself. Backup/snapshot software (this should be a feature of the file system), clipboard software (the OS clipboard should cover these needs), hot key software, window management, “anti cheat” (or really anything that needs to be a kernel module), antivirus, antimalware, the list goes on.

A properly architected system should have an operating system in charge of managing apps and resources and hardware, and apps which mind their own business. Cross-cutting “horizontal” stuff like what Acronis is doing here are reimplementing things your OS should be doing, and thus aren’t tested along with the OS itself, and are bound to have issues like these.

Or you run macOS and the first party stuff is so buggy (spotlight I’m looking at you) that you’re screwed either way.

oezi 267 days ago | | |

Explorer extensions such as the dlls which set the overlay icons used by Google Drive, OneDrive, Git, SVN, etc. are notoriously buggy. We want programmers to provide these extensions, but the documentation is bad and the interfaces are hard to program (measured by the amount of crashes caused by even a Google authored tool). I think Microsoft should have provided a test suite for these implementors to run or a certification program.

lmz 267 days ago | | |

Lock the system -> get sued (or at least abused) for locking out third party utility writers. Leave the system open for extension -> poorly written apps by others ruin your reputation. Add telemetry to detect what third parties are doing -> privacy complaints.

There's no winning here.

7bit 267 days ago | |

Try Veeam for Windows. It's free and it will not get more light weight that this. Tried Acronis last 15 years ago and they went from amazing to bloated and slow and terrible. I assume it still is.

kg 267 days ago | | |

Spent an hour trying to get Veeam to work and it can't connect to an agent installed on the same machine. Sigh. Thanks for the suggestion though.

Arch-TK 267 days ago | |

Software is crap everywhere. But on Linux I have the tools to quickly track down the bug and fix it, on windows I can maybe slowly track down the bug and attempt binary patching of proprietary software or hope and pray the developers will fix it after I beg them.

This is a privileged position as I am both a Linux expert, and an expert software engineer, and an expert troubleshooter/debugger. If you are not these things then I'm sure things don't work as well.

idiomat9000 267 days ago | |

Operating system kernels should be modal, with a performance mode, a maintenance mode, a server mode and a developer mode. Program cache eviction depending on the programs mode of belonging in the machines situation.

mook 267 days ago |

Speculation, since I don't have the relevant software:

- the Windows icon cache is dependent on the display configuration (in particular, bit depth, and maybe pixel density)

- unplugging a monitor therefore flushes the icon cache

- the relevant file is an explorer icon overlay handler (which is identified by the registry key one needs to mitigate the problem)

- therefore flushing the icon cache causes a mass refresh

- the handler is probably looking to see if some True Image process is running †

- maybe it needs that to show different icon overlays?

- it's probably also doing that for every icon (i.e. the caching isn't happening across different icons, at least when the cache is flushed)

† in hind sight, this is better done via a named pipe or mutex or something

(Edit: added line breaks)

andreareina 267 days ago | |

Further speculation: there's something that's working in n² time

fathermarz 267 days ago |

This is becoming known the software supply chain space, but Acronis software tampers with backups and lots of companies are dropping it. “True Image” couldn’t be further from the truth. Recommend Clonezilla for sure.

orbital-decay 267 days ago | |

>Acronis software tampers with backups

In what way?

fathermarz 267 days ago | | |

When creating a backup, it attempts to condense sectors through some non-standard proprietary algorithm that shifts addressing and tries to roll it up. It wouldn’t be an issue if when restored it reversed this process, but it does not and leaves holes, gaps, or new sectors altogether in the reconstruction. Because of the spaghetti nature of this, the user can only use their software unless they want to reverse engineer the patented algo.

magicalhippo 267 days ago | | |

And, does it do so when you ask it to do a sector-based backup, ie an image, rather than a file-based backup.

The latter is the default since it's more flexible and requires less space when doing incremental backups AFAIK, but obviously isn't a true image.

pixl97 267 days ago |

>These files are also missing Product Name, Company Name, and Product Version in the ETW fields and much of this information is also missing from the sigcheck output.

I have no clue what vendors do this, especially MS themselves. Has the exact look of files that were installed by a virus.

ronsor 267 days ago | |

This is wrong. Malware authors put in a lot more work to make their software look legitimate, even including valid version information and digital signatures.

saagarjha 267 days ago | | |

Hmm, but if I was a malware author and legitimate software couldn't bother to put in version information maybe I should also not do that so I can blend in better.

kiriberty 267 days ago |

There software was so bad. I once had purchased a license. But when I started using, in the setting there was a possible workflow that if you click that button/menu option it will reset the license key. It happened (because I am curious right?) then my copy was no longer valid. I called their customer support and that person treated my like a criminal trying to lie all the way. So explained numerous times and sent them purchase copies and screenshots. After some attempts they issues a new license key (they might have verified their stupid mistake). On top of that, the version I bought did not have a way to delete option. I had to purchase that feature. This was long time ago and I no longer use this sh*.

topspin 267 days ago | |

I use to license Acronis as well and I share your view of the software quality. Since then I moved on to Macrium Reflect: better UI, faster, fewer glitches. It paid off about 9 months ago when a 1TB 970 Pro NVMe failed. Flawlessly restored in about 50 minutes.

noisem4ker 267 days ago | | |

I can also attest that Acronis is crap and Macrium software is very high quality. Too bad, though, Reflect Free edition has been discontinued and the paid edition now locks you in a yearly subscription.

thr0w 267 days ago |

I don't think anyone who's run Acronis would be surprised at this. The whole thing feels clunky. I had a weird issue where it would crash my Dante audio setup every time it kicked off a backup.

stevenicr 265 days ago |

I saw this on a previous windows machine - only discovered it when having lots of things running and the system bogged down, did the ctrl-alt-delete and looked for everything using cpu.. strange to see acronis using cpu cycles when it was never set to backup anything.

I had installed it and planned to use it, but had not set it up - and yet it was using cpu daily.

stopped using.

ocdtrekkie 267 days ago |

I buy Acronis True Image but I certainly don't install it. I use it for cloning operations mostly.

I would agree with another commenter Veeam's free offering is a great backup choice as long as you have a computer you want to backup separate from your computer doing the backing up.