Windows Home Server Data Corruption - Latest Info.

[FireTech]

Windows Home Server Data Corruption Fix Status: No official fix until at least June 2008!


http://support.microsoft.com/kb/946676/en-us

http://blogs.technet.com/homes...date-on-kb-946676.aspx

http://mswhs.com/2008/03/10/data-corruption-issue-qa/

 

[bsobel]

Originally posted by: FireTech
Windows Home Server Data Corruption Fix Status: No official fix until at least June 2008!


http://support.microsoft.com/kb/946676/en-us

http://blogs.technet.com/homes...date-on-kb-946676.aspx

http://mswhs.com/2008/03/10/data-corruption-issue-qa/

Yea, home server is dead to me until this is fixed. Its a complete non-starter as far as I am concerned.

 

[loup garou]

This is incredibly disappointing. I've been running WHS for testing on a box in the office to test, but I just can't bring myself to buy a box for home use until this is worked out. I haven't been affected by the problem yet in the office, but it's a real confidence-killer.

 

[JackBurton]

Yeah, I think I'll stick with either Windows 2003 R2 or Windows 2008. Sure WHS is cheaper, but I like my data UNcorrupted.

 

[BD2003]

Yeah, its a bit absurd. Their storage system is so convoluted its ridiculous - there has to have been a simpler way to balance storage that could have prevented some of these problems.

 

[bsobel]

Originally posted by: BD2003
Yeah, its a bit absurd. Their storage system is so convoluted its ridiculous - there has to have been a simpler way to balance storage that could have prevented some of these problems.

Naa, the design is fairly ingenious, they just had a bug and a nasty one at that. So they are off to figure out how to fix it. WHS will probably be ready for prime time once done, for now Im staying hands off...

 

[RebateMonger]

It's a shame that this problem exists. MS Windows Home Server is one of those products that meets a major unfilled need, and this is a big disappointment. The good news is that MS still insists that the Backup function of WHS works as designed.

I guess this is one more example of why you can't completely trust online drive redundancy (whether it's RAID or Microsoft's Drive Extender technology) to completely protect your data.

 

[stash]

It definitely doesn't affect the backups, since they are not replicated, so DEMigrator never comes into play. And depending on your point of view, that can either be a good or bad thing

 

[RebateMonger]

Based up the latest information from Microsoft, it appears that a "safe" way to use WHS today is to install a single (large) hard drive in the WHS. The "bug" doesn't occur if WHS only has a single hard drive. Run the automated client backups onto the Server and also share your company's data files from the Server.

Keep ongoing backups of the Server on one or more external hard drives. Backups of the Server need to be done anyway, so this isn't an "abnormal" precaution. You can't trust important data to ANY single hard drive or redundant array, anyway.

Following the above practice should allow the safe use of WHS while waiting for a final "fix". Once all is fixed, you can toss on all those "junk" drives for extra storage.

 

[BD2003]

Originally posted by: bsobel

Originally posted by: BD2003
Yeah, its a bit absurd. Their storage system is so convoluted its ridiculous - there has to have been a simpler way to balance storage that could have prevented some of these problems.

Naa, the design is fairly ingenious, they just had a bug and a nasty one at that. So they are off to figure out how to fix it. WHS will probably be ready for prime time once done, for now Im staying hands off...

Yeah, I suppose so, but I had serious performance issues (not to mention this bug) with the WHS RC.

I'd really love it if they could have integrated it better with the desktop - move user folders to it instead of having separate local and server, but then again, with so many apps having the propensity to hijack the "documents" folder for their own purposes, its probably better this way.

 

[ViRGE]

Wow, I knew MS was in deep trouble with this specific bug, but I didn't realize it was this bad. I guess the rumors were true, they effectively have to rewrite the Drive Extender tech from scratch using a new data management algorithm; that's just about as bad as things can get. I'm fortunate, I don't have any corruption issues since I wasn't editing data on my WHS box in the first place, but this effectively kills WHS for now, I don't think anyone could recommend it or products based on it until MS fixes this bug.

The whole OS is slowly turning in to an abomination that is just a dry-run for WHS v2. We've all known WHS v1 will have a very short shelf life before WHS v2 comes out as a much better product (Longhorn kernel for prioritized I/O for DEmigrator and Media Center support, among other things) but I'm starting to wonder if it wouldn't just be better for MS to stop selling WHS v1 and wait until WHS v2 is ready late this year or early next. I like my WHS box, but this is painful.

 

[jalaram]

I ordered the WHS evaluation kit last week because I wanted to research using SageTV on it. I might still able to try it, but with only 1 drive.

Is the single drive recommendation for the whole system or just the shared Drive Extender? Could I put the WHS/Apps on one drive and just have the other drive as the shared drive?

 

[BD2003]

Originally posted by: ViRGE
Wow, I knew MS was in deep trouble with this specific bug, but I didn't realize it was this bad. I guess the rumors were true, they effectively have to rewrite the Drive Extender tech from scratch using a new data management algorithm; that's just about as bad as things can get. I'm fortunate, I don't have any corruption issues since I wasn't editing data on my WHS box in the first place, but this effectively kills WHS for now, I don't think anyone could recommend it or products based on it until MS fixes this bug.

The whole OS is slowly turning in to an abomination that is just a dry-run for WHS v2. We've all known WHS v1 will have a very short self life before WHS v2 comes out as a much better product (Longhorn kernel for prioritized I/O for DEmigrator and Media Center support, among other things) but I'm starting to wonder if it wouldn't just be better for MS to stop selling WHS v1 and wait until WHS v2 is ready late this year or early next. I like my WHS box, but this is painful.

Yeah, not having I/O prioritization is a bit ridiculous on a server that feels its appropriate to run a full chkdsk on all your drives every 6 hours, as well as seemingly shift data around at random for no apparent reason. No x64 support for backups was also a dealbreaker for me.

I'm just really hoping they go all out and integrate it with your network (roaming accounts) rather than having it be little more than an expensive NAT with backup capabilities.

 

[stash]

Originally posted by: jalaram
I ordered the WHS evaluation kit last week because I wanted to research using SageTV on it. I might still able to try it, but with only 1 drive.

Is the single drive recommendation for the whole system or just the shared Drive Extender? Could I put the WHS/Apps on one drive and just have the other drive as the shared drive?

The problem can occur in any system with multiple drives. It doesn't matter if duplication is enabled for a particular share or not. If you only have one drive, you cannot hit this issue.

The problem also only occurs when you read and write to certain types of files. Simply reading those files will not cause this issue. Copying the files to another machine, editing them locally and then copying them back will not cause this issue.

The issue is serious, but the main reason it has taken so long is because it took a very long time to actually reproduce the error. Now the main cause of delay is for testing the hell out of the fix.

64-bit client support is in PP1, which will be out in the first half of 08.

 

[ViRGE]

Originally posted by: stash

Originally posted by: jalaram
I ordered the WHS evaluation kit last week because I wanted to research using SageTV on it. I might still able to try it, but with only 1 drive.

Is the single drive recommendation for the whole system or just the shared Drive Extender? Could I put the WHS/Apps on one drive and just have the other drive as the shared drive?

The problem can occur in any system with multiple drives. It doesn't matter if duplication is enabled for a particular share or not. If you only have one drive, you cannot hit this issue.

The problem also only occurs when you read and write to certain types of files. Simply reading those files will not cause this issue. Copying the files to another machine, editing them locally and then copying them back will not cause this issue.

The issue is serious, but the main reason it has taken so long is because it took a very long time to actually reproduce the error. Now the main cause of delay is for testing the hell out of the fix.

64-bit client support is in PP1, which will be out in the first half of 08.

Unless something changes, PP1 will not be released until the corruption fix is released first.

 

[stash]

Unless something changes, PP1 will not be released until the corruption fix is released first.

I'm not sure where you're getting that from...

 

[ViRGE]

Originally posted by: stash

Unless something changes, PP1 will not be released until the corruption fix is released first.

I'm not sure where you're getting that from...

From Microsoft. They stopped work on PP1 to work on the corruption fix. They made it known way back when the corruption issue was first reported that PP1 would not be released until after they resolved the corruption issue.

 

[stash]

The release date for PP1 has not changed. It is still H1 2008.

(and yes I realize that June is within that period)

 

[ViRGE]

Originally posted by: stash
The release date for PP1 has not changed. It is still H1 2008.

(and yes I realize that June is within that period)

When you have a 6 month window, that gives you a lot of leeway to change things without technically changing things.

Anyhow, MS is considering releasing PP1 before the corruption fix. Apparently it was just about finished before they mothballed it. They're taking community input right now, and the community is clearly in favor of getting it out sooner. Microsoft complying with these requests would be the "something changes" scenario. The blog post also makes the issue clear that the current plan is to wait for the corruption fix, no matter what the current state of PP1 is.

 

[Scarpozzi]

I'm basically pushing all my data to FreeNAS and mapping drives to it. Network drives are a touch slow, but the hardware doesn't have to be anything spectacular and it does software RAID.

www.freenas.org

Setup isn't too shabby and it runs CIFS, FTP, iSCSI, and NFS.

I meant to add, with iSCSI support, you can grab the Microsoft iSCSI intiator and install it to map drives. If you can run a separate cat5 to it and put it on a VLAN with jumbo frames enabled, it's pretty fast. I've gotten this to work on a few ISO volumes I made and it's been flawless...it certainly is a step faster than mapping a CIFS share.

 

[Homerboy]

I'm perfectly content and happy with WHS. I haven't had a single corrupt file yet (since early betas) and over 1.8TB of storage (nearly full I'm embarrassed to say).
Though I will admit I don't do much, if any editing DIRECTLY off the server itself.

Couldn't you add 1 HDD into the WHS and not add it to the pool as your "UNEDITED DRIVE" and once you are done working with whatever files you are working, move them to their proper final resting space within the drive pool? Would fix the problem until June at least.

 

[Ares2600]

I'm having trouble understanding how any real software engineer can be bemoaning a race condition. In my experience they're not only some of the hardest to find but also some of the most common because they're the easiest to commit. Jumping to the conclusion that they need to re-write everything seems like assuming the worst case scenario. The sort example they give is kind of unfair. MOST race conditions are a simple matter of properly locking the resource that's being contended for. That's how they TEACH you what a race condition is, in fact.

I don't envy the guy who owns the code that this is in. Screw-ups in critical code are always pretty ugly whether they're honest mistakes or not. As I see it, however, there are ways to circumvent and the specific cases in which it occurs is known. That's 90% of the battle. Weigh your ability to work within these constraints against the other benefits you get from WHS. Make your decision accordingly.

Let's not have a software funeral here, and I personally think these articles are extremely dramatic. Bugs happen. At least this one isn't wired into millions of chips worldwide.

Edit: I personally wouldn't be putting my company data on Windows 'Home' Server, so if someone gets their business data corrupted I have no sympathy. Though to be fair I'd rather lose some work data than my track day pictures any day.

 

[WT]

WooPs ! I just bought a WHS yesterday with an extra 500gb drive for free, so could I set the drive to mirror the boot drive, or will it just dynamically add the drive (at which point I run the risk of data corruption?). I don't have to add the extra HD, but I'm not going to create a problem when all I need to do is show patience and await the eventual fix.

 

[ViRGE]

I'd just leave the extra drive alone for now, unless you need the space.

 

[WT]

Will do, ViRGE. I participated in the WHS beta and, altho I had less PCs at the time, found that tons of my stuff was redundant from PC to PC (ex: 15gb of music on each PC). I was genuinely impressed that the backups were so small compared to what I assumed they would be. The OfficeMax deal seemed to me to be a good deal, as the 470 could be simply changed to a 475 with the drive added.
I didn't do my research on the WHS issues until after ordering, but I truly think it is an addressable issue that will be resolved. Whats more important is getting my recorded media moved to a streaming server rather than stored on the HTPC.