Jump to content
Covecube Inc.

JazzMan

Members
  • Content Count

    61
  • Joined

  • Last visited

  • Days Won

    2

Posts posted by JazzMan

  1. 43 minutes ago, gtaus said:

    I have 2 4-bay MediaSonic ProBoxes and another 9 external USB HDDs all connected via USB 3.0 to my DrivePool server.

    [...]

    It's an option to consider if you continue to have problems with your TB3 JBOD box.

    I didn't name a brand, but the MediaSonic was the USB 3.0 JBOD I was having trouble with.  I got rid of that and built a small ITX PC with internal drives. That was about 5 years ago. I agree eSATA was more stable, but isn't an option in my current setup.  I don't have any experience with Thunderbolt (JBOD or otherwise) so was looking for some input in that area.  Glad your USB 3.x setup is working for you.

  2. Can anyone comment on how well DP would work with a Win 10 laptop and a 4 bay TB3 JBOD?  I've used DP for quie a while, but had most success with internal drives.  I did have an external USB 3.0 JBOD some years ago but it was rather flaky, often dropping the pool under sustained load (say copying a few hundred MB in one operation), which would then (as I remember) lead to needing to reboot the computer (I don't think re-plugging the USB cable or power cycling the JBOD was usually enough) and DP wanting to rescan the pool before I could use it again.  I blamed that mostly on the JBOD hardware/firmware though, and not DP.  Will TB3 provide a more stable pool, or am I likely to face similar issues?

  3. Sorry, I haven't been working on that system in a while. When I left it it was still a sporadic issue. I never got to the bottom of why the system would hang, nor why when that happened DP felt it needed to re-scan all 150TB. I would suggest you collect the logs Chris asks for above and submit a support ticket specific to your situation.

  4. In the wiki  it mentions rebooting after turning the key combination on.  You should also mention that "After you're done, you can disable the manual system crash key combination by double clicking on bugcheck_off.reg." you also need to reboot for the change to take effect. 

  5. Occasionally my server does not shut down completely.  It gets to a black screen but the power never actually goes off without holding the power button.  When this happens, on reboot DP goes through the lengthy re-measuring/checking of the pool(s) again.  I have not determined the root cause of why the system does not shut down cleanly;  I don't see anything obvious in the logs, and don''t make any software updates I can relate to this;  Windows Update is turned off.  Is there something I can do before shutdown to ensure DP gets stopped cleanly?  Can I stop the Stablebit service then shut down?  Would that help?  I'm guessing not?  Is it the file system itself that is marking that the pool needs re-measured.

    Win 7 x64 SP1
    DP 2.2.0.852

  6. It isn't really about the login account(s), it is about the owner and permissions that have been assigned to each folder/file.  NTFS tracks this as part of the file system.  Some users and security groups are built-in to Windows with predefined Security IDs which are the same on all machines (such as Everyone), other users and groups have SIDs derived from the machine they're defined on.

     

    By default the owner and permissions NTFS assigns are at the group level and for built-in groups.  If the old files were always on a single-user system where that user was in the Administrators group you should be fine.  But, if any of the pool/files were exposed on a network share, multiple-account computer, copied/moved via a removable NTFS disk onto the current system, etc. then they may have stray unexpected permissions.  Some software, e.g. where security is important, may also create files with special permissions.

     

    I found NTFS Permissions Reporter (http://www.cjwdev.com/Software/NtfsReports/Info.html) useful in tracking down files/folders that were not set with the default permissions and inheritance.

  7. Where did you get that case cause 45 drives doesn't seem to have it on sale by itself.

     

    I bought the Q30 including the mobo, raid card and power supply from them.  They list a few sample configurations on their site, and have a configurator for more choices, but when I dealt with them they were super flexible in working with each clients' specific needs.  The configurator can be a starting point but then you can order off the menu.  In my use case I wanted a different CPU than what they had on the configuraator; they could not source the specific one I asked for but understood my requirements and suggested an appropriate alternative they could source.

     

    You can check out these articles if you are not familiar with storage pods;

    https://www.extremetech.com/computing/178757-how-to-build-your-own-180tb-raid6-storage-array-for-9305

    https://www.backblaze.com/blog/open-source-data-storage-server/

    or search the web for something like "backblaze storage pod" and maybe add "open source" if you really are interested in just acquiring a case from a metal shop and not all the internals.

     

    For some history as I understand it;

    45Drives was an early supplier to BackBlaze but is no longer.  I think BackBlaze is up to version 6 or higher of the pod.  Around version 4 or so, BackBlaze switched to BackupPods.com and possibly others as suppliers of the pods.  However the pod designs are open source and can be bought and put together from a parts list if one desires.

     

    I needed something that arrived working and was not a server build learning experience, so I bought a system assembled and tested sans drives.  I then just dropped my drives in, installed Win 7 and DP and was basically ready to go.  Most drives were moved from some other computers, one of which had been running DP, and I added a few blank drives, so I did have to move some files around to get them into the new pool and duplicated properly. 

     

    http://www.45drives.com/

    https://www.backuppods.com/

    http://protocase.com/products/electronic-enclosures/45drives-storage-pods.php

     

    My perhaps faulty recollection was that 45Drives wouldn't necessarily sell you all the nuts and bolts and individual parts one by one like at Radio Shack of old but would be very flexible in putting together (or leaving out) the major electronic components that go into a server to a clients' needs, and that BackupBods was the opposite on either ends of that;.they would either sell you a completely built working machine with all the drives and little customization, or they would sell you all the nuts and bolts and individual parts a la carte to assemble yourself.

     

    I'd contact both companies and explain what you are looking for. (metal case, metal case with backplanes, etc.)  Protocase is the metal fab shop for 45Drives, so if you really are just interested in the chassis and not the backplanes, etc, you could contact them.

     

    This info may have changed in the past 1.5 years.

     

    Cheers,

    David

  8. OK, I see that now.

     

    But when I open the UI, how does it decide which pool to display the graphical info for first?

     

    I think that is what threw me off.  it always seems to default to displaying the graphical info for the second of my two pools on the popup list (it may be the first or last created or something)

     

    I guess I would expect it to not just sort the popup list but also either default to displaying the detailed pool info the first item in the popup list (my preference), or the last one displayed when the UI was closed.

  9. Read and write speeds probably aren't that great, especially from the HighPoint card, but luckily most of the time I don't need them to be.

     

    In a quick unscientific test right now copying a 9GB file between two drives (mount points outside the pool) averaged 130MB/s.  Copying that file into the pool (eg, from one of the 10TB drives into the 142 TB pool) averaged about 90 MB/s.  I've seen some tasks run a lot slower though, like 30MB/s.

     

    I'm not sure what my expectations should be in this area and haven't been geeky enough to learn to benchmark and twiddle, but it certainly isn't performing at 600MB/s headline speed of the 6Gbps SATA.

     

    Perhaps I'll look into that sometime, as I would like the task that measures the pool organization of the large pool to run more quickly, as it takes a half a day or something and leaves the system very sluggish while it is happening.

  10. This is the machine I use Stablebit DrivePool on.  It is not strictly a "server" as it gets used for workstation tasks as well and thus has a desktop, not server, OS.

     

    Case: 45Drives Storinator Q30
    OS: Windows 7 Pro 64-Bit
    CPU: Intel Xeon E3 1275v2
    MoBo: SuperMicro X95CM
    RAM: 2x8GB ECC
    GFX: NVidia GeForce GT 710
    PSU: Corsair HX750i
    OS Drive: SanDisk X300 128GB SSD
    Storage Pools:
      142 TB:
        10x 4TB (WD40EZRX)
        14x 6TB (WD60EZRX)

        4x 8TB (WD80EFZX)

        ~ About 3.5 million files (7 million with x2 duplication)

      18.2 TB:

        2x 10TB (ST10000VN0004)

         ~ About 155,500 files (311,000 with x2 duplication)

      * Pool sizes as reported by DP.

        Drive sizes as marketed.
        Both pools configured with x2 full pool duplication.

    Misc Storage: 750 GB (WD7500BPVT)

    HDD Controller card: HighPoint Rocket 750

     

    post-174-0-15759400-1505948266_thumb.jpg

  11. Sorry, is this sort by name an option in beta *.840+ or do I need to take an action to sort them?

     

    I am running beta *.852 and my pools are not sorted by either volume label or drive letter.

  12. Yeah, I know the risks you outlined with dynamic disks, but my backup software does support them and I have tested restoring the system from the backup.

     

    If your products don't support dynamic disks, I'd suggest testing for that and throwing a message, rather than just crashing.  I did run the troubleshooter and submitted the diagnostics,.

     

    Cheers.

  13. I installed the 32-bit beta in a Win 7 VM and gave it a test drive there so have a bit better idea how it is supposed to work now..

     

    After some light testing I converted the boot disk of the VM to Dynamic and rebooted the VM.

     

    CloudDrive broke as you said.

     

    I don't really see why Basic or Dynamic host disks would make a difference.

     

    However, this spanned volume was a quick fix for running out of space probably before I had DrivePool.  I'll work on redoing that.

     

    Thanks,

    David

  14. Win7 sp1 x64.

     

    Downloaded, purchased, and installed CloudDrive 1.0.2.929 β x64.  Rebooted although it did not prompt me.

     

    In the UI I connected to Google Drive.  That seemed to work, The status at the bottom is "OK".  There is an empty folder in 'My Drive' called 'Stablebit Drivepool'.

     

    Now I'm stuck though.  What am I supposed to do next?

     

    If I click "Create" next to the account under Google Drive, the UI just disappears.  I'm not sure if it is crashing.  I guess it is as the UI process disappears from task manager Processes tab.

     

    If I try to Connect a new Local Disk;

     - the Drive combo box is empty

     - If I click the ellipses button and select a (empty) folder on either my C: or D: drive I get "Data for a cloud drive cannot be stored at this location".

     

    This is a laptop with two SSDs, See Disk Management screenshot attached.

    - One SSD is in the built-in bay, the other is in a caddy in the DVD bay.

    - They are created as Dymanic disks and the D: drive letter/partition is a spanned volume across the two SSDs.

     

    I will try other providers but so far I'm batting 0 for 2.

     

    Please advise for each of these what I am doing wrong or info you need to troubleshoot.

     

    Thanks,

    David

     

     

    post-174-0-28290600-1505774314_thumb.png

  15. OK, it took some effort but I seem now to have both pools' duplication consistent.

     

    I still think there is some bug in the beta around the triggering for when to remeasure the pool and/or the actual logic that the duplication is not consistent.

     

    Here is what I did after fixing all the owner and ACL issues I could find as described above.

     

    - set all the drives of pool G offline
    - Did not touch Pool D and it was Green/OK as far as duplication consistency
    - stopped the drivepool service
    - uninstalled 2.1.1.561 from control panel
    - rebooted

    - installed 2.2.0.833 Beta
    - Received the "unable to connect to service" dialog (although the service is started)
    - stopped the DrivePool service
    - cleaned out the ProgramData and AppData directories for the program as described somewhere in this forum or Wiki as this seems to be a know/documented issue with 2.2 and/or upgrading
    - Restarted the drivepool service
    - Opened the DrivePool UI and Activated my Key
    - Pool D showed as Duplication inconsistent  [WHY?]

    - Re-measured pool D
    - Pool D now shows Green/OK [if the pool was inconsistent why did I have to manually re-measure it? and why did it miraculously become consistent with no change on disk?]

    - set all the drives of pool G online
    - Reopened the drivepool UI.  Drive G showed as "Measuring"
    - When done measuring, pool was "duplication inconsistent"
    - Re-measured pool G
     - Pool G then showed Green/OK

     

    - Rebooted cleanly.  System shut down and restarted without forcing any programs to close or BSODs or whatever

     

    - Upon reboot, logged in and opened the DP UI.

    - Drive D was OK/Green

    - However, Pool G was "Measuring..." again [WHY????]

    -  After being Measured for the Umteenth time Pool G shows Green again.

     

    - Rebooted

    - This time Pool G seemed to stay consistent across the reboot and was OK/Green when I logged in to the DP UI after the reboot.

     

    Thanks,

    David

  16. SYSTEM should have had access to all the folders/files.  I never had problems accessing the files, and i haven't noticed it missing.  "Users" was missing from some folders/files. 

     

    I'm less concerned about how the ACLs got out of whack as I understand things happen.  I never moved the whole pool, but it is possible some of those files were moved or copied into the pool from a portable NTFS drive in such a way as the ACLs were also carried over from the source system.  Or some such thing.

     

    Not to be a broken record, because you are always very helpful Chris, but if this is the root cause of the 'duplication inconsistent' or BSODs then DP should log something like "SYSTEM cannot access file \\pool\path\file.ext" not just say "Duplication Inconsistent".  Nothing in the thread above, or any other thread I have seen on the forum, or article I have found in the wiki, confirms for me that file permissions are even a thing that can cause the "Duplication Inconsistent" warning/error.  This tack is something I started on my own trying to find a root cause.

     

    I might take the pooled disks offline, try the beta upgrade one more time, bring the small pool back online, and see how it goes.  Overall I'm pleased with the DP product (I've had a few licenses for 4+ years), just not so much with this "Duplication Inconsistent" message and behavior.

     

    Cheers.

  17. Thanks Chris,

     

    I did some more looking in NTFS Permissions Reporter (on the native NTFS PoolPart mount points, not the combined pool) and found that there were some (maybe a few hundred) files and folders where it reported something to the effect that it could not translate the SID for the Owner to a user account.  (sorry I don't have the exact error any more).  I used the Windows TakeOwn command to correct that.  I set the owner to the Administrators group as that seems to be the default of how all the other files/folders were owned.

     

    This was done while DP 2.1 is the installed version.  I have not tried to re-upgrade to see if 2.2.0.x behaves better now.  .

     

    Again though, DP needs to report better on what inconsistencies it sees in the pool, not just that there are inconsistencies. 

  18. If the pool is consistent at the start of this process, will the pool be consistent after you bring the drives online, or will DP need to re-measure the pool after all the drives are back online?

  19. I just want to add that after fixing this as above I tried to re-upgrade to various beta versions, as 2.2.0.651 had been working, but that version, along with *.823 and *.830 all don't seem stable now on my system.  Even after deleting the recycle bin on each pool it is created on, I get a BSOD usually the first time I go to re-measure a pool after the post-install measure chugs away and finishes.

     

    It is really confusing/frustrating as the only change I made (intentionally at least) was to swap 2 4TB drives to 10TB in one pool.The second pool shouldn't be going inconsistent.  I seem to be the only one reporting this, but there seems to maybe be some problem with the Recycle Bin replication and/or handling of 10TB drives in the 2.2 betas.

  20. Is it possible to reorder the way the pools appear in the UI and/or the order in which they are measured?

     

    At the moment my large G Pool is listed first in the UI and is re-measured first whenever is considered inconsistent.

     

    I'd like my smaller D pool to be first in the UI and "Next Pool" would take me to the G pool.  D is more important, would re-measure quicker, and D comes before G in the alphabet.

     

    Thanks,

    David

×
×
  • Create New...