On ReFS

[brutalizer]

Here is a MS developer talking about why Windows lags behind Linux and other OSes. He says about the new touted ReFS which MS adverties as the best thing since sliced bread:
http://blog.zorinaq.com/?e=74
"Oh god, the NTFS code is a purple opium-fueled Victorian horror novel that uses global recursive locks and SEH for flow control. Let's write ReFs instead. (And hey, let's start by copying and pasting the NTFS source code and removing half the features! Then let's add checksums, because checksums are cool, right, and now with checksums we're just as good as ZFS? Right? And who needs quotas anyway?)"

 

[ochadd]

Funny how that writeup just reaffirms my take on Windows in general. Flashy neon lights that make the casual user/observer get excited but puts more bullshit between you and the needed results.

 

[Rectal Prolapse]

If you read the update you'll see the guy saying "NTFS isn't as bad as I said it was". To be fair, you could do a lot worse than NTFS hehehe.

 

[DPI]

Funny how that writeup just reaffirms my take on Windows in general. Flashy neon lights that make the casual user/observer get excited but puts more bullshit between you and the needed results.

Time to quit the LSD

 

[brutalizer]

The text only confirms my doubts on ReFS. MS thinks that you just need to add some checksums, because they are cool. But by default, metadata is not checksummed. How do you expect to get a safe solution by not checksumming everything? I suspect that if some researchers examined the data corruption detection abilities of ReFS, they will find it lacking.

And the write performance in storage spaces is abysmal (are they using ReFS? How are they related to ReFS?).

 

[bigdogchris]

I've always wondered why open source puts a lot of effort into improving file systems on a regular basis, where Microsoft goes decades between re-writes.

 

[danswartz]

Captive audience?

 

[x509]

I've always wondered why open source puts a lot of effort into improving file systems on a regular basis, where Microsoft goes decades between re-writes.

So lemme guess here. (No, I'm not trolling. )

You guys don't seem to think that ReFS would be a complete match, feature for feature, with same or better performance, than any of the various filesystem alternatives that are a staple of the discussions here.

 

[SirMaster]

And the write performance in storage spaces is abysmal (are they using ReFS? How are they related to ReFS?).

Storage spaces doesn't automatically use ReFS. ReFS is only available on Server 2012 while Storage Spaces runs also on Windows 8. Storage Spaces on Server 2012 can use ReFS, but on Windows 8 Storage Spaces uses NTFS.

 

[bigdogchris]

You guys don't seem to think that ReFS would be a complete match, feature for feature, with same or better performance, than any of the various filesystem alternatives that are a staple of the discussions here.

I don't know since I've never used ReFS. I was just commenting that Microsoft doesn't update their file system very often compared to open source. What I have read on ReFS shows that it has more features (of course) but that the performance is weaker than NTFS. But I'm sure when it came out NTFS was shown to be slower than Fat32 as well. It just takes time.

 

[madrebel]

You guys don't seem to think that ReFS would be a complete match, feature for feature, with same or better performance, than any of the various filesystem alternatives that are a staple of the discussions here.

currently it isn't. can it be? sure, when though?

ReFS needs 2 years of non mission critical run time before I would even consider it for use in mission critical workloads.