Don't count on being able to directly modify most of the spy/mal/ad-ware or its controls, especially in win11, without a fight - it has gotten increasingly tougher to take system control back from the mother ship since 22H2 and beyond... You will have to get tools and/or learn commands to take control of files and reg-keys, and even then you have to piecemeal most of the stuff security wise due to the mickeysoft "fuck you" level of security they place on all of their shit-ware and controls thereof. Many of the "privacy" tools/scripts used successfully on 21H1/H2 and below are now useless on 22Hx and beyond with this new level of "fuck you" security and the mickeysoft tightening of the crap-ware noose around most unsuspecting users necks by hardening the implementation of the garbage they load onto those systems. The larger issues with most of this is that many who run these unmodified stock systems have no clue as to either how to combat the bullshit laid out by mickeysoft or that they are simply unwitting pawns of the mickeysoft corporate testing/advert and malware departments.

As far as ReFS goes, the fact that it is unable to be implemented on a system drive is a tell (to me at least) that it is not ready for prime time for my machines anyway. The system drive is what takes a lot of the beating in a typical windows system so why there is no implementation of their wunderkind file-system for the system drive seems suspect in my eyes... To be perfectly honest after 20+ years of use I have still yet to push to the widely published boundaries of NTFS to the limits. NTFS is not a perfect system by any means, but it is really the only choice for general computing over any other FS's offered on mickeysoft OS's.

"ReFS is integrated with the Storage Spaces feature. If you set up a mirrored Storage Space using ReFS, Windows can easily detect file system corruption and automatically repair problems by copying the alternate copy of the data on another drive."

This to me is simply putting a band-aid on a problem that has been a part of windoze for nearly 30 years. Get rid of the corruption in the first place or simply mark "bad areas" of the storage medium as BAD and never to be used. Chkdsk would be the likely candidate that should do this but it doesn't ever seem to work that way and hasn't since the beginning of windows. You have to get third party tools to accomplish that task. There is nothing that any software is going to do when bad/malfucntioning hardware is involved other than replacing the bad hardware.

"If ReFS detects corrupted data and doesn't have an alternate copy it can restore from, the file system can immediately remove the corrupted data from the drive. It doesn't require you reboot your system or take the drive offline, as NTFS does."

Sure, just remove the corrupted data. The issues with this to me are twofold. The "corrupted data" shouldn't be there in the first place as in it never should have been "corrupted" - second - this software is now taxing the host hardware even with more useless overhead to make up for inadequacies that have been there since day one.

"ReFS doesn't just check files for corruption when reading and writing them. An automated data integrity scanner regularly checks all files on the drive to identify and fix data corruption, too. It's an auto-correcting file system. You don't need to use chkdsk at all."

Sure you don't. Just pound that drive with as much data as you can and leave the "driving to microsoft"...
As far as operational speed improvement there only seems to be improved performance under certain "virtual machine" and/or hypervisor situations. Nothing else is really improved speed wise across the board operationally. The "encryption" and "file locker" capabilities are useless junk to me. The ONLY semi-redeeming qualities ReFS has over NTFS is the < 255 character file path, and the < 16 EB file limit. Those limits of which I would ever come close to pushing...

meh.. whatever.

 

Last edited: Oct 6, 2024

Hey,

Do you have a cool list of processes that can be disabled , that you could share ?
1000x thanks

 

For win10 22H2
- Anything adobe related
- Auto Time Zone Updater
- BitLocker Drive Encryption Service
- Connected User Experiences and Telemetry
- Data Usage
- Device Management Wireless Application Protocol (WAP) Push message Routing Service
- Diagnostic Policy (all)
- Dialog Blocking Service
- Downloaded Maps Manager
- Encrypting File System (EFS)
- Geolocation Service
- anything google related
- Hyper-V (all)
- Internet Connection Sharing (ICS)
- IP Helper
- Microsoft Defender Antivirus Service (all) (replace with more viable solution)
- Microsoft (R) Diagnostics Hub Standard Collector Service
- Microsoft Account Sign-in Assistant
- Microsoft App-V Client
- Microsoft Cloud Identity Service
- Microsoft Keyboard Filter
- Microsoft Store Install Service
- Natural Authentication
- Net.Tcp Port Sharing Service
- OpenSSH Authentication Agent
- Program Compatibility Assistant Service
- Remote Registry
- Retail Demo Service
- Routing and Remote Access
- Secondary Logon
- Shared PC Account Manager
- Storage Service
- SysMain
- Update Orchestrator Service
- User Experience Virtualization Service
- WAaS Service (aka update medic)
- Windows Biometric Service
- Windows Camera Frame Server
- Windows Connect Now
- Windows Defender Firewall (replace with more viable solution)
- Windows Insider Service
- Windows License Manager Service
- Windows Media Player Network Sharing Service
- Windows Mixed Reality OpenXR Service
- Windows Mobile Hotspot Service
- Windows Modules Installer
- Windows Perception Service (all)
- Windows PushToInstall Service
- Windows Remote Management (WS-Management)
- Windows Update
- Work Folders

A lot of this stuff will require taking control of regkeys and or renaming/deleting files/folders. A large amount of task-schduler junk which is associated with the stuff above should be disabled/deleted as well.

 

Last edited: Oct 8, 2024

Barge in? To what exactly? Your b.s. session?

If it's a useful thing I have no problem embracing or praising it. In this particular instance, no. I do not find ReFS especially useful at all, IMO. You can do whatever the fuck you want, just don't piss and moan when someone else expresses opinions that differ from your own.
that's all...

 

Sir you visibly shit your pants in trembling rage anytime anyone mentions Windows 11, and whether you like it or not it's a perfectly functioning operating system for the overwhelming majority of people that use it. You can't act like "you don't care" about what other people use when you repetitively drop multi-paragraphed dissertations, definitively pissing and moaning, on how every Windows OS released past Windows 7 is bad. Come on man. Lmao.

 

Naaaah. Of course they weren't hahah. I did the exact same thing you did with the exact same drives
Early adopter you see, the drives were in my possession a week before they hit the shelves on stores and all i knew about them was just their spec sheet. Soon as i learned they were unsuitable for raid 0 i ... did nothing hahaha. I already had an image of my base installation so i thought fuck it, let's push this thing and see what happens. After all they came with a 5 year warranty. The smartass i was, i thought nothing is mentioned anywhere officially by WD, i will take the piss on those drives 'till they die and then i will claim warranty lol.
Performance in raid 0 was better than any other hdd, the partition was hitting almost double iops than a single drive and about 250-280 mb/s sequential reads. Luckily enough, the drives lasted in raid 0 a couple of years and in the meantime the whole planet and me had moved to Win7. After that, in 2010 i dismantled that machine with the drives still working(!) and upgraded to the next gen VelociRaptor, 600gb i think it was, but no raid this time and since 2012 it has been strictly ssds as system drives for me. And probably all else who seek a combo of performance and reliability.
Cheers mate

 

This proves absolutely nothing my friend.
Data decay or bitrot can happen on any digital device. With storage devices it is mostly common to drives that are used only for read purposes and are of some age. It is highly unlikely to happen on a drive that is used for writes and reads constantly. If it does it would be because of "wear and tear", meaning the device reached its end of life.
The main thing here, is you can't prevent bitrot on "read only" drives without any form of data integrity checks. What you say "the corrupted data shouldn't be there in the first place", is practically impossible. Unless of course you switch to new storage drives like every 6 months, something that is out of the question for the vast majority of users and i suppose you as well.
Personally i have 3 Kontakt library drives just on my main machine. These drives might get re-written perhaps once every year or two. I save my own presets, multi layered combos and whatnot on a different drive. While i always backup to my small personal storage server, i have lost to corruption Kontakt (and other sample libraries) drives more than 10 times the last 20 years. Just so you know 6 of them drives were on raid 5 (3 each time, 2 for data one for parity) and the second time this happened, the ~2 days lasting array rebuild failed continuously for a week+ and after that i called it quits and formatted the whole thing to oblivion hehe.
The case is simple with bitrot: Most of the times, partition turns to raw and Win asks for initialization. And sometimes the drive(s) may turn to a complete brick.
Now if you haven't experienced the data death of such a drive, then you are either very lucky or an absolute wizard. Or simply lying. A good percentage of my clients have experienced their music/photos/movie/audio libs storage drives getting corrupted over time and some of my friends too. It just happens because all of us just store files accumulated through the years and the drives just sit there to be read and nothing else and you can't do much about it, other than have a fs that can ensure data integrity. In Windows there's nothing else than ReFS. But don't worry we 'll soon know if it's worth it. It might very well be shite and everything mentioned as good about it, is just on paper. That's why we 're here. To poke a stick up the technology's ass and see how deep it goes hehe. Oh and be sure i hold nothing against you personally, if there is a specific test you want me to run on ReFS just say so.
https://www.digitalmusicnews.com/2024/09/13/hard-drives-music-storage-dying-report/

Now this is funny. Because i never thought i 'd be defending MS lol. But man. Most of the third party tools that supposedly check disks just trigger the internal chkdsk function of Windows. The whole planet agrees there is no better alternative. If it's good for Acronis it's good for anyone in my book. Unless of course they are too MS's corporate partners in crime hehehe. There are some few tools that can also do a sector by sector check but still chkdsk is the best to check the file system's integrity. Which tool was it that you are using please?
Surely commands like chkdsk/f -- chkdsk/r -- chkdsk/x and chkdsk/f/r/x have to be used with caution and users are always advised to backup their data before proceeding with such a command, as removing/altering data blocks or whole sectors on a drive can be disastrous for the data's integrity. And speaking about NTFS/FAT32/exFat, the bitter truth of it all is, no 3rd party software guarantees that after a drive check leading to a "successful" error and/or bad sector repair etc etc, your crucial data will be intact 100%. If you know of such a tool please elaborate, i 'm all ears.
Cheers

 

Last edited: Oct 7, 2024

Without getting into details. Would reformatting my HDD drives to ReFS have any noticeable read/rewrite speed or integrity impact when dealing with Sample Libraries or Kontakt libraries?

 

With single volume HDDs it should have a slight impact in performance (10%) because the FS constantly generates checksums but it will ensure data integrity until the drive reaches its physical end of life. With SSDs the performance impact should be negligible, even SATA ones.
ReFS should be best suited for drives that consist of sample libraries of all sorts and are used just for read purposes.
I will soon check it out on a brand new machine, if you can wait about a week or ten days i will have posted my findings by then.
Cheers

 

Last edited: Oct 7, 2024

ReFS is great for a backup drive where files are written once and rarely change. It's not great for heavy I/O. That's the simplest way to look at it. So yes to storing your Kontakt sample libraries, no to your default Studio One song folder. But the benefits aren't significant enough that anybody should rush to move everything to ReFS. It's just something worth considering moving forward.

 

Whatever. You have your opinions I have what I have experienced over 20+ years of using winblows systems. I have heard self proclaimed "computer technicians" talk about some stuff you mention however I have never personally experienced it with any of my systems. Might be due to luck or it might be due to decent paring of devices into the hardware builds I have run over the years coupled with keeping usage within the limits of the hardware. Winblows has problems which were there at its inception and still linger in win11 - mickeysoft doesn't give two shits about it. Those of us who have used it for years know about them and simply work around them. Otherwise we'd all be using apple shit. The mickeysoft way is to simply throw more crappy ass code at a problem rather than fix the root cause. These days they are more concerned over how much garbage-ware they can load into a system to collect usage information on unsuspecting dupes. With every "update" session they scan your system for reverse engineering and "correct it for your own good" making well certain that their ad/spy/malware crap stays intact and working for them.

As far as chkdsk goes re: marking bad sectors/clusters I have had mixed results with it over the years. Regular backup imaging and judicious running of systems within their limits has proven non-problematic for my uses. I have also run many solid-state and HDD storage devices in many different systems and never experienced the "corruption" you speak of even close to the degree you specify. Perhaps the root cause behind your astronomical degree of data corruption is ReFS, or an improper or misconfigured implementation of it. I do know lots of folks who run these systems way past their design limits and expect the hardware to perform over the top with no resulting consequences which is foolish. They end up with problems such as data corruption, overheating issues, or hardware failures way before the OEM quoted mtbf... Either, as you said, you are particularly unlucky, or possibly over extending your hardware past its limits. I really could care less one way or the other.

So, I'll continue with my way of doing things which has served me well over the years... So we agree to disagree and we can leave it at that.

EDIT:
RE Hardware RAID: Comments saying hardware RAID isn't used in enterprise scenarios (or even home use) these days == complete poppy-cock. Either you haven't ever been in a data center or you have been talking to the severely misinformed. Hardware RAID IS used in MANY enterprise scenarios when speed, rock solid redundancy, and where offloading of storage processing from the mains is a requirement for the application. I believe a comment was floated saying something about RAID controllers aren't even available anymore and/or dont support but raid-0 or 1. Total horseshit. Just because hardware RAID controllers don't show up on your PC part picker doesn't mean they do not exist. Maybe do some research before making such misinformed comments.

 

Last edited: Oct 8, 2024

Like you can't read. ReFs is not on my menu or has ever been in my systems. This thread is about testing the thing lol. I dunno what you think you know with 20 yrs of using Windows but oh how sad you 're never satisfied with it, but i have been running a pro recording studio in 3 different iterations since 1994. Just ask and i 'll post photos anytime. And discography. I don't hide behind anonymity, it's not how i roll. And since 1984 (the year you were born probably?) i have tried all comp systems and OS's including Unix and Free BSD even godamn BeOS and Solaris lol. The thing you wrote in your first post about "heavy working environment" has always been my game. I 've been driving all sorts of top notch comp systems in my possession to "death from exhaustion", a good 25-26 yrs now. So add me to the list of the folks you know running their systems past their design limits hehehe. Yeah you know the foolish ones lol.
A piece of advice if permitted. There is test machines on one side and workhorse comps that you need stable on the other. But. You will never know what a peripheral or a cpu/gpu or even a mobo's VRM etc etc is actually capable of, until you drove it to its limits. It is what will teach you everything that the manuals don't include and what will actually make a good technician or even enthusiast tinkerer a better one. You can't know the real limit unless you exceed it. Btw, my QX6700 Core2 quad extreme @ 2,66 ghz from late 2006 still runs at 3.5ghz after 18 years. But yeah you know best man, foolish ones...

First of all you are rude. Manners maketh man mate.
Secondly, you are all talk no show, presenting no evidence but still swear at others for something you think you know young man.
Thirdly again i will advise to read carefully because you misread. But np i will say this, find a single desktop mobo from the last 5-6 years that supports raid 5. It's rhetorical because there isn't any.
Number four, you are missing the point entirely.
Dedicated controller or not, redundancy is still applied on a software level. Now explain to my humble ignorance what do i need the highest end Broadcom (LSI) 9660/9670 with 8gb of ram, in a comp with 128 cores, 1tb of ram and 128 pci-e gen5 lanes. It may be top notch hardware no doubt about it, but still it sits in there just like a fly in a cup of milk if you know what i mean. Alas it's not on par with the machine's spec not to mention its still pci-e gen 4.
But let's speak numbers. Wanna know how much of cpu/ram toll will have a 32 nvme drives soft raid implemented on such a system? 2-5% on the cpu depending the iops of the drives and about 32-64gb of ram. As negligible. And it will perform faster than the Broadcom LSI controller. Now i highly doubt you ever put your hands on such a raid controller device that i mentioned or server for that matter but i will still expect an answer even if it's only theoretical.
Lemme enlighten you in how evolving is the hardware raid market. This key study from Intel, Dell and Fujitsu expects a 4% increase in SAS RAID Controllers until 2030. Not just me, but anyone who's been around lately will tell you that with the amount of AI clusters piling up exponentially year after year and month after month, this is totally ridiculous considering a good part of the server/ws mobos in circulation include an embedded SAS RAID controller anyway.
https://www.linkedin.com/pulse/sas-...024-comprehensive-study-top-neal-yadav-xvmlf/
Lastly, please don't come back with insults and speaking without at least some type of hard evidence to justify your claims. My personal acceptance is one of FAT16 figures and my fault tolerance very limited with no redundancy whatsoever. You will leave me no choice but to press ignore, something that i have only done once in the past (@Foster) and i think i have been quite polite to you despite your negative tone. Not to mention humor is not your forte ?
Cheers
PS: Storage Review, the leading web authority in enterprise and consumer class drive reviews/reports, admit in their own Broadcom MegaRAID 9670 review from 2023 that they have been avoiding hardware raid cards for years. But yeah what do they know hehehe.
Read all about it here: https://www.storagereview.com/review/broadcom-megaraid-9670w-16i-raid-card-review

 

Last edited: Oct 8, 2024

hmm. I thought we were done. Well. Now I'm done.

 

After a couple of months of testing ReFS i have some info to share. First of all i tried ReFS with 3 types of drives, Pci-e ssds, sata desktop single volumes and enterprise hdds in soft raid 0 and 5. I will cut it short, because the results were somewhat unimpressive.
For single volumes, formatting in ReFS adds nothing and it actually subtracts about 10% of the drive's speed. It wouldn't be so bad if it didn't also somehow reduce the drive's responsiveness and this stands for both ssds and hdds.
For raid volumes with parity like raid 5 configs, it seems to be a good solution, as it performs similarly to ntfs and by definition it features a more robust data integrity. About this last one, i haven't been able to actually verify a better data integrity. All i can say, is it held on to its claims when it comes to parity volumes and it never gave me probs. Also, data transfers from one parity storage pool to another similar are greatly improved over ntfs, i measured about 30% better speed.
For striped volumes again it was worse than ntfs, especially with ssds ( i tried with 2x Samsung 990 PRO and 3x) it had about the same drop in performance as with plain single volumes only a bit worse due to the extra layer of the raid header.
Test machine was a Ryzen 7900X/64gb ram/Asus Hero X670E. System drive was a Samsung 990 Pro 1tb.
Test drives were three Samsung 990 Pro 4tb, 3 Seagate EXOS hdds 12tb, 1 Barracuda Pro hdd 8tb, 1 WD Black hdd 4 tb, and 3 WD Red Pro hdds 8tb i got on loan from a friend.
Lastly, Storage Spaces is better now, my previous experience with it on Win Server 2019 and Win 10 Enterprise was bad. It has since been improved, in both functionality and system integrity. I dunno if i would recommend ReFS for a NAS machine, but if someone wants a nas running Windows for their own reasons, it is quite good for parity raid volumes (raid 5/6).
Cheers
PS: It could also be useful in a workstation environment (Xeon-W series/Threadripper) in a machine with lots of storage drives where data integrity is key and the user needs parity volumes.

 

Last edited: Dec 27, 2024 at 10:42 AM

I think the limitation of shared PCIe bandwidth is the culprit

 

Are you talking about access time or similar? That would suck big time.

 

Yes. Ever so slightly, but if you have experience with the particular drives i mentioned you would notice it right away. It's in the nature of the FS, it constantly calculates security bits. Takes a toll on single volumes. At first i was hoping it wouldn't be apparent but on single volumes it just is and you don't need to run CrystalDiskMark to tell you so, it's that obvious.
From what i saw, speedwise it's only good for parity volumes. Those who want an "inexpensive" large Kontakt partition based on enterprise/nas hdds can make good use of it for instance. This particular scenario works well, like the 3x 12tb Exos drives: set one drive for parity and two for normal storage and you end up with a partition of about 22tb that can do ~350mb/sec transfers and almost 70% more iops than the same single drive on ntfs. The same partition on ntfs wouldn't exceed 300mb/sec mostly. I ended up keeping the raid 5 array partition but for another purpose, mainly for the client's local backup who also backs up to external drive as well.
Cheers mate