Christopher (Drashna)

Doesn't matter. Specifically, the file system check is a read-only chkdsk scan, which is fine with moving files when it occurs.

there are a couple of things to do here... First, run a memory test. Any time you BSOD, it's a good idea. Just in case it's faulty memory. After that, download "WhoCrashed" from Resplended Software, or "BlueScreenViewer". Then use whichever program to look at the cause of the BSOD. That should give you an indication of what caused the issue. Or you could email me a link of the "C:\Windows\MEMORY.DMP" file and I can take a look. I wish my A8 idled at that temperature..... But I do use it more aggressively. It is usually at 25% load all the time, because of Blue Iris (IP Camera software, and CPU intensive). Also idling at ~40C bothers me. It should be in the mid to low 30's if the CPU is idle, IMO.

it's not necessary, but it would be the simplest way, yes. Once the folders are removed, restarting the service will cause DrivePool to re-check those drives and see that they're not part of a pool (anymore). And what I mean by "removed" is that the contents are removed from the PoolPart folders and those folders are removed. This effectively removes the disk from a pool. Otherwise, yes, what you posted will work just fine:

Disconnect the pooled drives. All of them. This (hopefully) should let you boot into the system without running the disk check. Once it's done that, use the "show disconnected devices" "feature" (?). http://support.microsoft.com/kb/315539 Open up Device Manager (run "devmgmt.msc") and find the "Covecube Virtual Disk" drive. (Should be under "disk drives"). Right click on it and select "Uninstall". DO NOT select the "delete drive files" option. Reconnect the disks and reboot the system. If this doesn't help, unseed the pool (on another system). To "unseed" it, basically move the contents of the hidden "PoolPart" folders on each disk to the root of the disk. Then delete the now empty PoolPart folders. Reconnect these folders to the system, and boot up. Then "seed" the pool. http://wiki.covecube.com/StableBit_DrivePool_Q4142489

I agree about the APUs. They run hot and eat a lot of power. The only advantage is that they may support ECC RAM out of the box. As for transcoding... Yes. Absolutely certain. Most of my files are MKV. Plex will stream these directly to the Home Theater systems, so that uses very little CPU. However, both the Plex website and Subsonic require transcoding to a browser compatible format. And Plex does DLNA, IIRC (but I hate DLNA). Awesome! And that board officially supports the Xeon CPU as well. So god call. Just make sure you take the wife out to a nice dinner, too!

Yeah, I'm not a fan of spinning the drives down, in part because my server is fairly active all the time. Additionally, we had solar installed in June, so I'm not as worried about electricity costs. But I absolutely understand wanting to have the disks spin down. And specifically, the issue here is more than likely that StableBit Scanner are constantly "pinging" the disk, when it queries the SMART data. Open up the Scanner Settings option, and click on the "General" tab. Set the work window, if it's not already. Set the time to occur when you're going to be active on the server (so the disks will be awake anyways). Then click on the SMART tab, and check the "Only query during the work window or when scanning". This will only query during that work window, and should allow the disks to sleep better. As for the disk control, that means that your disks firmware doesn't allow you to configure these settings. Some will show the settings, but won't "take". Additionally, for StableBit DrivePool, any time something accesses the pool, it may wake up the disks. For the first access after a boot, it will wake up the disks. Afterwards, it will cache the file location and wake up only the disk the files reside on. However, programs such as Windows Search (or other indexers) can cause the disks to wake up more frequently. A test here is to stop the DrivePool service (and the Scanner Service) and see if that curbs the behavior. You can help pinpoint the cause of the disks being woken up this way. As for the wear and tear, that really depends on your use case. Entirely. As I said, I use my server a lot. I also share my files with friends and family. I also use a few other tools to identify duplicate files and monitor file integrity. Additionally, Plex periodically scans the entire library to update the files and add new files. As does Subsonic. For my use case, leaving the disks on all the time makes sense for me. But that's me. That's not you, and may not be your use case. As for the WD Red's, they are a very good option. For sleeping or always on as well. You'll get good performance out of them, as well.

I'll see about getting the upload limit increased (or feel free to use dropbox or the like and link the image). Not sure what to tell you. If it comes back, let me know. But hopefully, it stays working properly. Also, just to make sure here: the client back up is a single instance storage engine. That means that additional backups don't take up additional space.

To clarify, a single volume can ONLY be part of one pool. You MUST remove it from the pool (or remove the PoolPart folder), to add it to an existing pool. Additionally, if you connect the disk from the second pool to a system, it will see it as a different pool and create a new, second pool. As for resetting the settings, rebooting the server will restart the service. Sorry for the confusion. But to clarify further, you don't need to reboot the system during this entire process. Once the last disk in a pool is "removed" (manually or via the UI), the service will unmount and remove the pool. That is why you need to start the service back up, in this process.

Beaker, You're very welcome. This was a topic that was very important to answer, IMO. As for the Disk Details, that would be why. That information is populated by the "Identify" command. Since we avoid querying that data for these drives, you won't see it (unless you enable that "GoodIdentifyOverride" option). And yeah, they should increment the firmware version number. That's the ENTIRE point of releasing a patch like this.... And I'm glad to hear it! As for the slated programs, I can't give ETAs, but we do want to get to them sooner, rather than later. As for the BitFlock ID, both the stand-alone program and StableBit Scanner use BitFlock. The stand alone app lets you create a "nest" that lists your drives, and uploads information about them (such as the SMART data). You can access and share this data. Additionally, both products store a number of "rules" on how to interpret the SMART data, as well as which drives to "ignore" for stuff like the "identify" command. It also includes a lot more information about the disks, and the controller's they're attached to, which we can use to "fix" problems, or even blacklist devices. It's a big part of StableBit Scanner.

Ugh, the files are tiny and hard to read.... It looks like the tasks are fine. Just out of curiousity, if you try to run the restore, do you see both sets? If so.... try disabling the backup, reboot the server and client, and then re-enable. As for the Add-In install... this is for "pushing" client add-ins to the connected PCs. Such as for WHS2011 Media Collector, or LightsOut. Both of which have a client side installers. This task pings the server, checks and then will install them, if needed.

@hobester: Make sure that you have TWO (or more) SSD drives. If you don't, then duplicates will be written to the "archive" disks which may defeat the point of having the balancer plugin installed. Also, if you're using the "Ordered Placement" option in the SSD Optimizer, then there are a couple of settings you need to tweak to get this to behavior correctly: UNCHECK "File placement rules respect real-time file placement limits set by the balancing plug-ins." CHECK "Balancing plug-ins respect file placement rules." UNCHECK "Unless the drive is being emptied." However, for the files affected by the rules, this may break the SSD Optimizer (as it leverages the Real Time Placement limits to make sure it works). I've also flagged this for Alex (the developer) to get his input on how to handle this. But could you clarify which balancers are enabled, which order, and the specific settings?

You are very welcome. And yeah, set a calendar date to remind you, if you need.

Ah, yup, the SSD Optimizer will DEFINITELY do that. It's only to ensure that files are placed on the "SSD" drives first. Also, to be clear here, if you're using the SSD Optimizer, then you don't want ot use the Ordered File Placement balancer on the same pool, as it causes issue. Instead, use the "Ordered File Placement" section in the SSD Optimizer.

Okay, just to clarify here (for you and the future)... While this is absolutely terrifying (IMO), Alex (the developer) is already aware of the issue, and has code in StableBit Scanner to detect and "handle" these drives in a 100% safe manner. What I mean, is that we DO NOT sent the "identify" command to these drives, regardless of the firmware version. Since this is what cases the destabilization and data loss, it means that "we" can safely query the drive without compromising your data. So you can safely enable the SMART queries for the drive. However, that means you may not have as much information in the disk details section (such as firmware version) for these drives. I'm sure you're 100% fine with that, though. However, you can override the behavior, if you're so inclined. http://wiki.covecube.com/StableBit_Scanner_Advanced_Settings Add your model to the "GoodIdentifyOverride" setting.

Wow, that's a pretty scary bug! As for the drives, then yes, on the "Disk settings" page, I would recommend check the "Do not query SMART" option for these drives. However, if there is no important data on the disks, then I would say attempt to reflash the disks, just to be sure. Grab a permanent marker and write "flashed" or something similar on them, so you can be certain in the future. However, we do have the option to only query SMART data during the "work window" for StableBit Scanner. If you open the Scanner Settings, click on the "SMART" tab. Check the "Only query during the work window or if scanning" option. Additionally, could you get the BitFlock ID for the system? (open the "SMART details" for one of the disks, check the "submit to bitflock" option, and post the ID once it's generated it).

Iammis, That is very unusual, as normally, CHKDSK fails on the pool drive immediately (as there is no actual media to check). However, normally, when scanning during boot, there is an option to skip the check. Usually by hitting the "Esc" key. This happens usually right after the system starts to boot from Windows.

Yes, the M1015 is a "clone". Or more specifically, and OEM rebrand with custom firmware. Also, the Norco RPC-4220 case is a very nice one. I have one, in fact. As for the SAS expander card, they're basically a SAS "switch". You plug in one or more cables, and you can attach many more. They function in the same way that a "Port Multiplier" chip works in an external enclosure. But with a LOT more drives, and potentially much higher throughput. They're not exactly cheap (but neither are the controllers). http://www.newegg.com/Product/Product.aspx?Item=N82E16816117207 In theory, SAS (what the LSI and highpoint cards are using) can address up to 255 drives (I think). That means, you can get far more than the 8 drives that the card can support by default. As for compatibility, I'm not sure. In this case, you'd definitely want to contact the manufacturer and check beforehand. And forgive me, I'm still rather new to the SAS and related tech, as I've only recently really started getting into it. However, I enjoy learning, and I find the enterprise tech very fun. As for iSCSI, that is definitely pretty neat. And a great way to supply storage to systems without physical disks. Though, what's really neat, is that if you do it right, you can boot windows off of an iSCSI volume.

Which balancers are you using? If you're using the SSD Optimizer or the Ordered File Placement balancer, then this is normal. This is how it prevents DrivePool from placing files on the other drives, and forces it to use specific drives.

No, definitely not. Actually, I'm using PerfectDisk Server on my Server 2012R2 Essentials box, and have for a while. Works fine with the pool. Though, I'd recommend the "Performance Aggressive" option for hte pooled drives. That way, commonly used files are quicker, where less used files may be "further out" on the spindles and take a little more time to access.

The Core i3 CPUs are great. While not power houses, for most "home usage", they should be fantastic. I run Plex and Subsonic, both of which transcode video (and audio). I also run VMs (2-3 at any given time). So the Xeon e3 I'm running is very useful. CPU Benchmark is a fantastic site for comparing CPUs. And half the TDP means less power and less heat. Worth the money, IMO.

Yeah, the utility changes the share location, or shares the folders. It's well tested, so it should definitely work. As for removing DrivePool... there isn't an easy answer there. Since there is a good chance that files may be scattered between the different disks.... it would be messy at best. I can cover it, but .... it would be basically moving a lot of files around, and hoping you have enough disk space (or you could use the "File Placement Rules" to force it initially, and then "unseed"). And no, the utility only works for the pooled folders. However, if you're adventurous, you can "hack" it by using Registry Editor. Or... create a new shared folder and move the contents into that folder Well, the util page contains a description about each of the processes. However, we'd like to add a series of Youtube videos about the util and other aspects of the products. (if a picture speaks a thousand words, a video speaks millions) However, it's a time consuming and potentially expensive process. That works too. And that's part of how the troubleshooter util works. Yup. Well from the Dashboard, at least. You can also use the "windows" share manager to mess with this. Or you can use Registry editor to do this... The "shuffling" is what I mentioned above. As for the service, that's fine. Mainly we instruct you to stop the service, so that it prevents the service from attempting to duplicate the files on the pool, or from balancing files. It's not strictly necessary, but to prevent potential complications.

Also, if you removed the drive, one should be hidden and one unhidden. The unhidden directory is the old one. That should be the one to delete. As for the "other", you may want to manually remeasure the pool (under Pool Options, in v2.1)

Yeah, it definitely shouldn't be an issue. However, if you're using ESXi, then you may want to check out this thread: http://community.covecube.com/index.php?/topic/131-how-to-get-smart-data-passed-on-from-esxi-51-host/ SHould allow you to successfully pass on the SMART data to the VM.

GijsJan, You absolutely can safely remove these folders. Specifically, these are Windows Update temp files. They're placed on the disk with the most free space. But they are safe to delete once the updates have completed. As for preventing the from coming back? There really isn't a good way, unfortunately.

Yeah, it should be able to get the SMART data from the controller. I'm using an IBM ServeRAID m1015 card, which uses the LSI SAS2008 chipset (similar to the LSI9240-8i card). However, some have reported that you need to use the "Unsafe" DirectIo option. If this is the case, then you need to do this: http://wiki.covecube.com/StableBit_Scanner_Advanced_Settings Find the "DirectIo" section, and check the "Unsafe" option. Then reboot the server (or just restart the "StableBit Scanner Service"). It should then see the SMART data. And that's a nice board! Don't forget to setup IPMI on it.