• Content Count

  • Joined

  • Last visited

Posts posted by eweitzman

  1. @omnicorp



    I’m using the USB passthrough trick where you set a USB hard drive to offline and attach it directly to a VM in Hyper-V


    How does unraid verify the drive's "signature" against the key file without being a flash drive?



    - Eric

  2. @mattie112 I've created a bootable VHDX with a copy of my flash drive's contents. unraid's boot menu displays, and unraid will start to boot, but it fails while creating the virtual file system with these messages:


    /dev/root: Can't open blockdev
    VFS: Cannot open root device "(null)" or unknown-block(0,0): error -6
    Please append a correct "root=" boot option: here are the available partitions:
    [no partitions are listed]
    Kernel panic - no syncing: VFS: Unable to mount root fs on unknown-block(0,0)


    How were you able to boot an unraid guest under hyper-v?



    - Eric

  3. Hi.


    I built a new system around a Supermicro M12SWA-TF motherboard and threadripper pro CPU. It's been running headless for a week now with video and keyboard on the BMC controller to access IPMI and unraid's boot screen. I just  installed an RTX 3060 PCIe video card and am trying to get the system working with it.


    The UEFI bios is set to send VGA to the PCIe video card. The system and unraid work with the video card: during boot, diagnostics come to the monitor connected to the 3060 and the login prompt displays. When I choose to boot unraid in GUI mode, however, after the boot diagnostics display,  when I'd otherwise expect the login prompt, the screen clears and displays a cursor at the top left corner of the monitor, as if unraid's GUI can't talk the the 3060. If this is a driver issue, how would I resolve it? I'm not familiar with how video drivers are installed from the shell on a headless linux setup, and if they're available through apt-get or similar, how would I make this persist the flash without having to reinstall on every boot?


    Second, assume this gets resolved, and also assume I get the video passthru to a windows VM working. I've already installed a windows VM and can access it from the VNC viewer in a browser or through RDP when the system is headless (without a gui boot). How would I set things up so unraid would start the VM such that it displays directly to the monitor via the 3060? In other words, without any user interaction, this machine should appear to boot into windows on the attached display.



    - Eric

  4. As I documented in this general support thread, unraid 6.9.1 requires the root account to have a password for many functions to work. I've been able to run unraid with no root password (and without the friction this causes) for over a decade. Nothing in my circumstances or practices has required changing this now.


    A root password is required by

    • ssh
    • samba connections from windows 10
    • web UI


    Root passwords are not required by

    • telnet
    • console access


    Please either revert to the behavior of previous unraid builds that don't require a root password, or make this configurable.



    - Eric


  5. @John_M - Thanks for taking the time to point this out in the release notes. (They've moved to https://wiki.unraid.net/Manual/Release_Notes/Unraid_OS_6.9.0#SSH_Improvements now.) This explains what I've discovered re ssh requiring that root have a password, but doesn't explain either why this is now required nor why other avenues of access (telnet, web UI, samba) do or don't require it. I'll have to find out how to register a complaint with management.



    - Eric

  6. I have no password on the unraid root account.


    rsh/ssh prompt for the password and will not log in when the enter key is pressed. This is new behavior in 6.9.1: I have been using ssh for root without a password successfully like forever.


    telnet connects without requiring a password.


    The GUI itself opens without requiring a password.


    The web GUI console opens without requiring a password.


    Any suggestions?



    - Eric

  7. The os returns "No such file or directory" when I try to put unraid to sleep by this command:


    echo 3 > /proc/acpi/sleep


    The directory is there and appears to be writable. The file "wakeup" is there. I can cd to /proc/acpi. I get the same error trying to create any file in /proc/acpi. User is root: sudo shows no difference.


    This is behaving as if the directory or partition is read-only.


    root@unraid:/proc/acpi# pwd
    root@unraid:/proc/acpi# whoami
    root@unraid:/proc/acpi# echo 3 > sleep
    bash: sleep: No such file or directory
    root@unraid:/proc/acpi# touch x
    touch: cannot touch 'x': No such file or directory
    root@unraid:/proc/acpi# ls -ld ../acpi
    drwxrwxrwx 3 root root 0 Apr  3 12:27 ../acpi/


    I have the Dynamix S3 Sleep plugin installed. The sleep menu item in the GUI header does not cause the system to sleep.


    Any suggestions?



    - Eric

  8. Hi. I'm a longtime developer, new to development on slackware. Discovered pkgtools is the package manager. But it seems that the gui program pkgtool is missing from unraid 4.7.2. The other pkg commands {remove, install, etc}pkg are here and running. So I pulled pkgtool out of the lastest package distibution at http://ftp.slackware.com/pub/slackware/slackware64-current/slackware64/a/pkgtools-15.0-noarch-28.txz, put it in /sbin with the other *pkg files, chmod +x, installed dialog, et voila, I have pkgtool for the moment -- at least until the next reboot.


    pkgtools-15.0-no-arch-23 is installed on my machine. I haven't found an archive of all older versions of slackware packages yet, so I can't confirm whether pkgtools-15.0-noarch-23 was missing pkgtool or not, but the source code for slackware 14.2 (2016) does have it, and the web is littered with references to it. So I expect it should be present.


    This has caused me to wonder, not only if Lime has removed this package for some reason (anyone know?) and I shouldn't install the latest pkgtools, but how can I know if installing some random package will interfere with unraid itself?



    - Eric

  9. On 3/5/2018 at 7:54 PM, Frank1940 said:

    If you acknowledge the error, you should not get another warning UNLESS you have a  new error.  As @johnnie.black says even an occasional one is not a problem.  In fact, you should never lose data because of a CRC error.  Once the error is detected, the data will simply be resent until it received correctly.  (I still have not quite figured out why they decided to highlight what is a minor problem in virtually all cases.)  


    I skipped 6.4.1 and went to 6.5 and have two drives with crc errors. Would you mind describing how to acknowledge them? I see the crc counts in the smart listing on the affected drives' pages highlighted in orange, but there are no buttons, etc, to acknowledge the error.



    - Eric

  10. Will updating a drive's firmware cause any data loss?


    If not, is it safe to update the firmware if the drive is part of the array? That is, will any portions of the disk be changed (such as a signature) that would invalidate parity?


    In case the answer depends on the drive/manufacturer, I'd like to update a WD Red WD60EFRX from version 68MYMN0 to 68MYMN1.



    - Eric


  11. Interesting advice, but has rfs really rotted such that it's no longer usable?


    This 6TB drive has 750GB free so the "nearly full" theory doesn't apply. I have several other 6TB drives that *are* nearly full but have never had rfs problems. In fact, this is the only time I've seen this problem in seven years of using rfswith unRAID.


    The array has been up and moving data around, including from/to this disk, for about 12 hours now. Mystery.



  12. A 6TB Red Pro drive with ~1000 hours has twice had write failures and reiserfs problems while running mover and once for no reason that's apparent to me. SMART short test shows no problems.


    About 2 weeks ago while running overnight, mover caused several thousand reiserfs errors and unraid red-balled the drive. reiserfsck --check informed me to run --rebuild-tree. This took a very long time (10+ hours?) with the array in maintenance mode. All looked well afterwards, until a day later under normal use, unraid red-balled the drive again but without reiserfs errors. I replaced cables and moved the drive to a different controller and used data-rebuild to bring the drive back online. 11M writes and 15 hours later, the array came up and worked normally. Then, when running the mover, it failed again with reiserfs errors.


    Is there any other way to further diagnose this problem besides running a long SMART check?



    - Eric



    reiser errors (mover).png


    reiser errors (reiserfsck).PNG


    reiser errors (SMART).PNG

  13. I noticed that I missed a "5" in one of the port mappings. It's now "55555". Restarted the docker, it required that I re-enter my license, but still can't access my /sync folder.


    This is the docker command now:

    root@localhost:# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d 
    -e TZ="America/Los_Angeles" 
    -e HOST_OS="unRAID" 
    -e "PUID"="99" 
    -e "PGID"="100" 
    -p 8888:8888/tcp 
    -p 55555:55555/tcp 
    -v "/mnt/disk1/ResilioSync":"/sync":rw linuxserver/resilio-sync


  14. I would like my /sync folder to be inside the array, not on the cache drive. To that end, I set /sync to /mnt/disk1/ResilioSync, created the directory, made sure user:group was 99:100. When I try to create a manual folder connection, enter a key pasted from sync on another machine, and use the path in the "Connect to folder" dialog as set by they share/key to /sync/docs (for example) and press "Connect", the dialog says "Sync Home does not have permission to access this folder."


    What am I doing wrong?



    - Eric





  15. When I first started using unraid, I bought into the notion that storage solutions like this could run on any ol' hardware you had lying around unused in a closet. unraid upped the ante by being able to use any ol' disks. While this was and still is all true, I discovered it also led to much frustration in terms of low speed, wildly varying transfer rates, and mysterious long random waits and freezes. In spite of these frustrations, it was perfect for what it did at a very low entry cost. I just had to temper my expectations and appreciate the job it did.


    A few months ago, I bought a beefy mb/cpu/ram combo from user @KeeWay. It's a gigabyte server mb with 32GB ram and a fast Xeon cpu, $1K of 3-4 year old technology for about $300.


    The new machine is a real pleasure. It is liberating to actually be able to work directly against files on the array without huge delays and freezes such as when editing metadata. I had basically given up on doing anything other that copying files to the array and playing them from the old machine. Of course, I would make minor ad hoc changes, but now unraid storage behaves almost as well as local disk on my desktop. There is still the occasional delay when creating files, but that is rare now instead of typical.


    I was in the process of coalescing many old smaller slower drives into 6TB drives in the months before I got the new hardware and finished that project recently. With the drives on two PCI-e SAS cards in this beefy machine, things like parity check regularly average about 100MB/s, peaking at 140MB/s.

    Next I'll see just how fast data can move from an SSD on my desktop to the array using a pair of 10G ethernet cards (ebay pulls). I might have upgrade the cache to SSD to fully exploit this.


    A whole level of frustration and compromise has gone away after graduating to better hardware.


    - Eric


  16. I just redid steps 10-13. I believe I must have pressed "Done" in the New Config page without pressing "Apply" first. Now, after pressing both buttons, every device on the main page has a blue square indicating "new device" in hover tips and the "Parity is already valid" checkbox is present.


    One suggestion for improvement: After checking the "Parity is already valid" checkbox, the warning up top beside the parity disk should change from "ALL DATA ON THIS DISK WILL BE ERASED WHEN ARRAY IS STARTED" to some other message indicating it will be preserved.


    The array has started with my old disk gone. Time to start a parity check to be sure.



  17. I've almost completed removing a drive using the method described here: https://lime-technology.com/wiki/index.php/Shrink_array#Alternate_Procedure_for_Linux_proficient_users. I am absolutely certain that I have unmounted, cleared, and unassigned the correct drive.


    A problem occurs in step 14, "Click the check box for Parity is already valid, make sure it is checked!" There is no such checkbox on the main  page.


    There is the following checkbox after I unassigned the zeroed out drive:



    Start will disable the missing disk and then bring the array on-line.
    Install a replacement disk as soon as possible.
    Yes I want to do this


    The server is running version 6.3.2.

  18. I order this drive and got the $25 off for using Paypal. The cost was $235 delivered - no shipping, no tax. What a deal!


    The drive is in a retail blister pack, the blister pack is in a 12"x15"x6" box with air-filled bags that are completely flat. The bags have small tears in them from the sharp edges of the blister pack. The blister pack itself has bent corners and one part is compressed. The drive has undergone lots of impacts from UPS and USPS handling.


    It may have been a good drive when shipped, but probably has been subjected to G forces out of spec. It's going back. What a disappointment.