Christopher (Drashna)

Sizzl, Thanks for posting this. We will look into it, and see about fixing it. And if this is for 2.x, if you use the "Options" button, does it work fine?

Efficently? Recommended? It really depends on what you mean, and what you want. Depends on preference. IF the disks are the same size, then it doesn't matter, you'll ge the same. But my recommendation is percentage. If the drive is marked as damaged by Scanner (if you have that installed as well), then it will attempt to migrate data off of it. You can configure it to do the same with SMART errors. And for either, you can set it to move unduplicated and/or duplicated files off the drive in either the damaged or SMART error state. Also, it can be set to avoid using drives that have exceeded the temperate threshold, as these drives are more likely to fail. However, adding a "health" rating, and using that to avoid using disks is a good idea, and I'll pass that onto Alex. ANd I'm glad that you love our software! As for more drives.... there are other options... pricy... but ..... I'd recommend any LSI chipset based card. They're expensive ($200 range), but 2x miniSAS SFF-8087 cables means 8 SATA HDDs without using expanders. .... That, or I blame a friend for this link, as I can actually use it (Norco RPC 4220, with 20 hotswap bays): http://www.newegg.com/Product/Product.aspx?Item=N82E16813157470 22 SATA Ports, and .... uses LSI chipset and an expander to accomplish this. And it's a haswell board. But I don't "need" it. Per se.

For the lockup, did it actually BSOD? If it did: http://wiki.covecube.com/StableBit_DrivePool_System_Crashes At least upload the memory dump via the Widget at the bottom of the link. However, if it didn't BSOD and it just hung, if it happens again: http://wiki.covecube.com/StableBit_DrivePool_System_Freeze

yes, that is the add-in for WHS, but I believe the version is 1.3.5.7572. And doug is right about this. Both work, but they are very different in the UI. But for WHS2011, I do recommend 1.3.

Actually, unless you have installed the "Disk Space Equalizer" balancer, it won't equalize the space across the disks. Well, that is unless you're adding new files. New files tend to be placed on the drives with the most space available, which will tend to equalize the disks eventually. But the "Disk Space Equalizer" balancer enforces this behavior.

You have Network IO Boost, and Read Caching enabled (if the files are duplicated), under Performance, right? If so, what is the speed like copying from the Pool to another disk (preferrably a SSD or something *very* fast)? Also, what about the speeds from the individual disks in question? I would post another couple of links to see if the network is the issue, but you're already basically confirmed that it isn't an issue. So could you also enable "tracing"? http://wiki.covecube.com/StableBit_DrivePool_2.x_Log_Collection

dscline, Unfortunately, the issue may be a simple as the drive changing the SMART data on us. Or it could be that the drive has triggered it's internal error correction and fixed some of the issues, causing the data to change, etc. Could you download the newest beta of Scanner? http://dl.covecube.com/ScannerWindows/beta/download/ http://dl.covecube.com/ScannerWhs2/beta/download/ It should be 2.5.0.2978. This version has a lot more detailed logging, and if it happens again, it could catch it. However, to enable this logging, you need to open up Scanner, and open the "Scanner Settings". There shoudl be a new "Enable advanced settings" checkbox on the general tab. CHeck that and close the window. In the Settings drop down list, there should be a new "Advanced Settings" box. Open that, and in the list it displays, find "SmartDiskInfo" and click on it. Then select the "Verbose" option and hit "OK". This will dump all the SMART data to the log files so we can see what is going on. However, this can cause the log to grow large, very fast. Also, if you restart the service, or reboot the system it will revert to the previous state (to prevent it from filling the disk with log files).

UnsafeDirectIo uses a number of methods that we have marked as "unsafe" because they can cause issues with the controller/drive/drivers, that can cause issues between causing the controller to restart, to blue screening the system. That is why they're "unsafe". And while it may be a small percentage of systems and a small chance, since Scanner queries this information periodically, we'd rather use the methods that we know are safe. This is from the guide on how to change the advanced settings, and is about the UnsafeDirectIo setting: If you still have questions, please do ask.

Okay, I'm pretty sure this is a WHS issue, but still... http://wiki.covecube.com/StableBit_DrivePool_2.x_Log_Collection Don't enable Tracing, just upload the logs. Also, could you find, "send to->compressed folder", the "C:\ProgramData\Microsoft\Windows Server\Logs" folder? Upload the logs to the same place as the Stablebit Logs (the link above has a Box.com upload widget), just reference this thread in the comments.

That's definitely not supposed to be that way. But could you go to the DrivePool tab real quick, and check the letter assigned to the pool? If that matches, make sure there are no "missing" disks here.

Precisely what otispresley has said. Specifically, the reason is that the installer doesn't use the "Windows Server Solution" installer (WSSX extension), so we don't use the normal methods to install DrivePool. However, if you uninstall DrivePool from the Control Panel, it will uninstall the dashboard tab, as well.

Okay, thanks. As for the pool disk, yes it should show up as 2TBs in disk management. That is definitely normal.

For the https://stablebit.com/Buy/DrivePool link? If so, it appears to be working fine for me. Have you modified your CAs on that system at all? That or may have been a temporary glitch.

Bob, That is very odd that it would do that at all. As for the formatting, you can't format the Pool drive at all. You'd want to recreated the pool. But if the pool is completely empty, and the disks that you are using for the pool are likewise completely empty. Then it may be a good idea to format those disks and then re-add them to the pool. Aside from that, before doing anything to the pool, could you enable trace logging and re-attempt the action that caused the error? http://wiki.covecube.com/StableBit_DrivePool_2.x_Log_Collection And you are definitely right, I'm seeing next to no information about this error.

Shane, you're absolutely right about DrivePool being a storage solution, and not a backup solution. It's meant to protect against drive failure. Not against corruption, deletion or the like (while it may help against corruption when checking duplication, that's not a major focus, as that's a disk health issue) This would be more of a question for Alex. We do store a lot of logging information by default (C:\ProgramData\StableBit DrivePool\Service\Log\Service\), and if you enable tracing, it shows exactly what the system is doing, file by file. But those logs are a lot more complicated to read. Though, we (well, Alex) could add a more verbose logging of what is going on during duplication and balancing. It definitely couldn't hurt, and may help with troubleshooting some issues.

Well, the Pool is "portable". As long as you move all the pooled disks over, once you have DrivePool installed on the new system, it will recognize the pooled disks and recreate the pool. Nothing that you need to do, and no moving over 20TBs of data.... just the disks. And the Pool is backwards and forwards compatible. So if you move the DrivePool 2.x, you will be fine, as well. Though, you will want to download and run the "WSS Troubleshooter" utility. Namely, run the "Restore DrivePool Shares" options. This will reshare the folders from the pool for you. http://wiki.covecube.com/StableBit_DrivePool_Utilities

Well, there is no "upgrade" option. It's a clean install or not at all. As for the pool, reinstall the OS, install DrivePool 2.x and then run the WSS Troubleshooter and use the "Restore DrivePool Shares" option: http://wiki.covecube.com/StableBit_DrivePool_Utilities But if lag is the issue you're having.... then I have a link for you: http://forum.wegotserved.com/index.php/topic/8335-before-you-post-media-stuttering-playback-issues-performance-irregularities/ Ignore the fact that the link is for WHSv1. All the same things still apply. Well, aside from the services stuff.

#2, Any system should support it as a non-system/boot disk. Aka as a data disk. It is only really important on the boot drive, where BIOS needs MBR, and UEFI needs GPT. And yeah, WHS2011 runs very well, And there is probably a reason why some people want to use it as a desktop OS... But a SSD really does help out. As for the file placement, there is the "Ordered file placement" balancer plugin, but that's not really what you want. It fills one disk at a time. But for these really important files, you could change the duplication count. There is a command line utility (dpcmd) that can allow you to set the duplication per folder or to a higher than x2 duplication (btw, DrivePool 2.x can do this through the UI natively) http://wiki.covecube.com/StableBit_DrivePool_Utilities But as for the technet/msdn keys, even if you're subscription expires, they will be stored and accessible. As my free acounts for being an MVP expired a year and a half ago... and I just checked a couple of days ago, and they're still there.... I can say that with certainty. For the rest of it, setting the duplication higher would be a good idea. As would using CrashPlan or SkyDrive or some other off-site backup solution. What happens if a surge kills your server (I mean, you should have a UPS, but things happen anyways), or if the location the server is at catches fire? It won't matter what drive it's on.

You are very welcome. I try to provide as much info as possible without presenting too much. Okay, so definitely "clean". DrivePool can add that in that state just fine. Because for the most part, the main reason you'd want to use one of the other is that GPT supports..... really large volumes. Like millions of TB sized volumes., IIRC. But if you want, you can initialize the disk as GPT, create a basic disk, and use a NTFS partition. DrivePool doesn't have any issue with that. However, we don't currently support Dynamic disks. Yes, DrivePool definitely assigns a letter. And you can change that letter, if you want. In fact, we recommend doing so, and to A or B, or something much higher up in the alphabet to avoid the disk letter from being bumped (Windows will do that some times when you add a new disk) Not a problem. Sounds good. But DrivePool is fine with moving disks around. Heck, the pool can be moved to a different system without much hassle (security permissions being the main thing) The files are stored in hidden "PoolPart.xxxx" folders. Everything in these folders will mirror how the pool looks. As for the pool itself, it's a virtual drive. It doesn't actually "exist" anywhere. All the IO operations (including listing the contents) are passed onto the disks themselves. (And the kernel tends to cache some of these requests in memory). If you want to back up the pool, you'd want to back up these folders. But remember, if they're larger than 2TBs, then Windows Server Backup cannot back them up. Also, information about the duplication status and some Pool related stuff is stored on the drives in the pool. However, most of the other settings (such as balancer configuration, activation information, etc) is stored on the system disk. So, basically, the pool really doesn't touch the system disk, unless you've added it to the pool (which isn't an option for WHS2011, actually). And talking about the OS drive, a small SSD like that is a great idea. In fact, my server runs with a 120GB SSD (I have ... a few things installed that need more space), as well as a drive dedicated to system roles (hyperV, Windows Deployment Servers) and download temp. None of the pooled data ever touches the system disk. And I have those drives backed up daily. And I'm glad that you've appreciated the thorough answers. And again, if you need any further clarification or have any other questions, don't hesitate to ask. And enjoy the New Years, as well! Regards

If they're "cleaned", as in uninitialized (such as running diskpart's clean command), the DrivePool will partition and format them. If the disks are just unformatted/RAW, I'm honestly not 100% sure. But it sounds like you're talking about "clean" disks. As for initializing them, it depends on the size. 3TB drives it formats as GPT, but less than 2TBs are formatted as MBR, IIRC. As for the WHS not supporting GPT, that's for client backups. I used SBS2011E for a while, which was near identical to WHS2011, and it used my 3TB drives as GPT with absolutely no issues. I haven't used BitTorrent sync, so I'm not sure. But it shouldn't have any issues with DrivePool, as far as I'm aware. Nope, DrivePool is file based, not sector based, so there should be absolutely no issues with the formatting of the disks. Either or. See above. You can have DrivePool installed without a pool without any issue. This is a misconception. The backup feature of WHS2011 does not support volumes larger than 2TBs. That means it can't back up 3TB drives or larger. This is a known limitation of the Windows Server Backup feature, that the dashboard uses. As I have said above, I used SBS2011E for a while, and had no issues. Well, I only backed up my system disk, and that worked fine. However, you cannot backup the pool directly. You would have to back up the pooled disks. I think I've answered everything here. If I haven't, or if you need clarification, don't hesitate to ask. Regards

Okay, that's good to know.

That's very odd. My IMB m1115 reads the smart data just fine.... Also flashed to IT mode. As for the way to enable it: http://wiki.covecube.com/StableBit_Scanner_Advanced_Settings Set "UnsafeDirectIo" to "True" and restart. Or use this file: https://www.dropbox.com/s/0krodbweqz1rrc4/Scanner.Service.exe.config

Well, then that seems to be a SMART reporting issue on the drives. Unfortunately, there isn't much us, or any other software can do about that. And if the disk did have 200 reallocated sectors and it's saying it's not now..... that's disconcerting, and I would recommend RMAing them, just in case. Better safe than sorry.

Lee, What specific features are you looking at? And if you don't like the more advanced features of Server 2012 (R2) Essentials, there is always this: http://www.tinkertry.com/how-to-make-windows-server-2012-r2-essentials-client-connector-install-behave-just-like-windows-home-server/

HyperV is much more simple. You just set the disk offline in disk management (or there is a way to set *all* new disks as offline, which is GREAT for hyperV), and then when you add storage to a VM, you just select the disk from the drop down underneath the "Physical Disk" option. The down side is that HyperV doesn't pass through the SMART data, while ESXi, RDM can.