Posts by RonO

    select pool

    Have you created your pool(s) in the KVM plugin?


    I have a basic pool set up of iso files and kvm files

    To do this go to the Pools tab


    click Add, Name = iso files, Type = Directory, Path = /path/to/directory/where/iso/files/reside or use the browse button to select a directory, click Save. These iso files are the images u use as installation cd/dvd's. see Optical Drive in Add VM (new vm) or Devices - Add Optical (existing vm)


    click Add, Name = kvm files, Type = Directory, Path = /path/to/directory/where/vm/image/files/are/saved or use the browse button to select a directory, click Save. This pool is where the vm images are saved to.


    Now when u create a new vm (VMs tab - Add) your Storage Volume - Volume will list available images to attach to ur vm, or when selecting Volume - Create new disk u can select where to save ur disk/volume/image/qcow2 file to (select kvm files pool). And Optical Drive will list all ur images/iso files from the "iso files" pool.


    Basic, simple and neat.

    I'm having a similar issue as OP.

    Been converting OS drives from other computers using a docking station (P2V) with

    sudo qemu-img convert -p -O qcow2 /dev/sdb '/mnt/OMV/nfs share/imagename.qcow2'


    This has worked great for legacy BIOS systems, 3 done so far.

    But now I'm having problems with a UEFI system drive, LinuxMint 18.3 installed on drive, partition 1 EFI boot, partition OS, partition 3 swap.

    Will only boot to UEFI shell.


    I can attach the imagename.qcow2 image to another vm and read the partitions contents.

    So it is not corrupt.


    No partitions mounted at conversion time (worked well for legacy conversion, no accidental reads or writes).

    Using OMV KVM plugin (by ryecoaaron ) to configure and run vm's.


    Question: Is there anything else / special I can set in the KVM plugin set up?

    I've tried Chipset i440FX & Q35, UEFI On & Off, Storage Volume Bus virtio & sata

    noVNC in the plugin shouldn't use port 6080

    My Krusader container is using port 6080 for novnc, as you'll see in ps aux | grep websockify below.

    The KVM plugin VM's are using your standard ports 8081 etc.

    Code
    root 6791 0.4 0.0 69124 11744 ? S 05:46 1:11 /usr/bin/python2.7 /bin/websockify -D --web=/usr/share/novnc/ 8081 localhost:5900
    root 6878 0.0 0.0 51252 9988 ? S 05:46 0:01 /usr/bin/python /usr/sbin/websockify --web /usr/share/webapps/novnc/ 6080 localhost:5900
    root 18712 0.0 0.0 49544 4864 ? S May12 0:07 /usr/bin/python /usr/sbin/websockify --web /usr/share/webapps/novnc/ 6080 localhost:5900
    root 22875 0.0 0.0 63336 4444 ? S May11 0:14 /usr/bin/python2.7 /bin/websockify -D --web=/usr/share/novnc/ 8081 localhost:5900
    root 22885 0.0 0.0 63304 10692 ? S May11 0:12 /usr/bin/python2.7 /bin/websockify -D --web=/usr/share/spice-html5/ 8091 localhost:5901
    root 30282 0.0 0.0 2392 764 ? S 09:59 0:00 sh -c (ps aux | grep websockify) 2>&1
    root 30283 0.0 0.0 2392 92 ? S 09:59 0:00 sh -c (ps aux | grep websockify) 2>&1
    root 30285 0.0 0.0 3220 892 ? S 09:59 0:00 grep websockify


    Implementing a PXE server in a docker container

    It shouldn't unless you are using host network on the container.

    The PXE server in a docker container is set to use the host network, but with it's own dnsamsq running within the container.

    I guess it's more research and testing.

    Thanks for your help ryecoaaron.

    Install 5.1.1 that is in the repo now.

    Question 1.

    Having an issue with the KVM plugin effecting other running novnc instances when starting (or Auto Starting) a VM in the KVM plugin.

    I can work around the issue by restarting the other novnc instance, it is Krusader file manager running in a docker container using novnc as a connection method, using port 6080, which also auto starts.

    Form an OMV reboot, it only occurs when the Krusader container is running before a kvm vm is started.

    Clicking the "Link" for the KVM vm will open Krusader with novnc,

    The KVM VM's are using the standard 8081 etc ports.

    This has started occurring since the 5.1.1 update.


    Question 2.

    Is LibVirtd related to the KVM plugin at all?

    Researching implementing a PXE server in a docker container that will use dnsmasq, concerned that the container instance of dnsmasq may interfere with the instances of dnsmasq running under LibVirtd.


    Thanks for all your great effort on the KVM plugin ryecoaaron it's working out to be an excellent addition to OMV

    ryecoaaron could you give some guidance please?


    I'm setting up linuxserver/netbootxyz docker image and dnsmasq DHCP in proxy mode

    I've been following this article

    https://wiki.jarylchng.com/boo…-has-locked-dhcp-settings

    But I noticed libvirt using dnsmasq in the processes.


    Can I add another configuration to dnsmasq without effecting the KVM plugin?


    The additional dnsmasq.conf is as follows

    port=0

    ##interface=eth0 ##not used, all 3 nics wanted for netbootxyz

    bind-dynamic

    log-dhcp

    dhcp-authoritative

    dhcp-range=192.168.1.0,proxy,255.255.255.0

    pxe-service=x86PC, "NETBOOT (BIOS)", "netboot.xyz.kpxe", 192.168.1.253

    pxe-service=X86-64_EFI, "NETBOOT (EFI)", "netboot.xyz.efi", 192.168.1.253

    I'm considering a project of setting up a pxe boot server on my OMV nas, by using the linuxserver/netbootxyz container.

    One of the boot images will be a virt-p2v environment, but that's down the track, I have to crawl before I can run.

    From reading articles and youtube stuff, I think I can use dnsmasq to achieve what I want.

    The reasoning for using dnsmasq is I have an ISP provided basic router, this has to remain as the DHCP server. It has NO nextserver or 66/67 instruction set.

    Can dnsmasq in OMV be configured to do this? Why I ask is that it seems a general rule, not to play around in the back-end of OMV, as the web UI may overwrite manual changes.

    If the OMV dnsmasq can be used for this purpose, how do I go about configuring it with the netbootxyz container ? or could someone point me in the direction of some good articles about this topic.


    The 2 clearest instructions I have found are (3 days of weeding through rubbish to get to this point)

    [HOWTO] PXE-Boot various systems

    https://wiki.fogproject.org/wi…le=ProxyDHCP_with_dnsmasq

    https://www.youtube.com/watch?v=U3RC20ANomk

    Show advanced options

    Thanks ryecoaaron

    I can't believe I have over looked that for so long, doh, note to self again, read the b#%@#&* screen.

    Is the way of adding a sound card to an existing VM via creating a new VM (with Audio On) and attaching the existing VM storage volume ?

    Or can it be added via the Edit XML method ?

    Feature requests @ryecoaaron


    1.

    Would it possible to store the Spice, VNC & noVNC port numbers with each vm?

    This would allow the AutoStart feature to be complete.

    Done a reboot of server test yesterday, the 1 gap in it all is the noVNC docker container is not running.

    The container Restart policies may have to be changed to "Unless Stopped", "Always" or "On Failure", the "Unless Stopped" option would be flexible for me.

    Fingers crossed this is not outside the scope of the KVM plugin.


    2.

    Option to turn auto refresh on or off

    I have found Auto refresh makes a selection from combobox list a task that needs to be done to quickly, I like to read through the available options before making a selection, but the auto refresh keeps closing the list on me.

    It is fine if the /etc/hosts file has the right value in it. :)

    Mine did not.

    # This file is auto-generated by openmediavault (https://www.openmediavault.org)

    # WARNING: Do not edit this file, your changes will get lost.

    127.0.0.1 localhost.localdomain localhost

    127.0.1.1 ????

    # The following lines are desirable for IPv6 capable hosts.

    ::1 ip6-localhost ip6-loopback

    fe00::0 ip6-localnet

    ff00::0 ip6-mcastprefix

    ff02::1 ip6-allnodes

    ff02::2 ip6-allrouters

    ff02::3 ip6-allhosts

    192.168.???.??? ????.local ????


    Copy of my omv hosts file

    omv placed the real ip record on the last line

    The container itself has to be able to resolve the OMV host's name. If you are using 1.1.1.1, that will not know the ip address for your hostname. Can you try changing your OMV system's dns server to your router?

    Found what was blocking noVNC, can't Force Secure connection only on OMV (System, General Settings, Enable SSL/TLS true, Force SSL/TLS true)

    Having just Enable SSL/TLS true, works ok, although it slows connection time and may not connect on 1st attempt.

    Once that was solved I done some tests with DNS settings, I found noVNC would still connect via hostname.local with any DNS server ip, tried router-dns-server ip and cloudflares 1.1.1.1, still worked.

    Although that may be from your recent changes to use "the host's IP address instead of the hostname"


    Can I ask for 1 feature request?

    Could "?resize=remote&&autoconnect=1" be added to the containers connection string?

    Complete string would be like http://hostname.local:8081/vnc…ize=remote&&autoconnect=1


    Many many thanks for the great plugin and the fast responses.


    PS Just tested connecting via VMC-Viewer and noVNC on pc's at the same time, all 3 can connect at the same time, this is handy for me as I do forget to close things, ole fella syndrome...

    I was trying to keep it simple

    Dooh, my bad, I was the one being simple and doing something wrong in my case...

    I had a pool added with sub-directories in it (virtualbox folder structure), the .vdi was in a sub-directory and I was selecting the directory name to Convert!

    Note to self, read the screen thoroughly.

    Thanks for the reply and the great work on this KVM plugin.


    I'm still unable to connect via noVNC, you mentioned DNS may be an issue, what DNS servers can I try?

    Currently using gatewayIP and 1.1.1.1

    All VM's configured to use your Default network, also tried a macvtap bridge network.

    OMV server has a static assigned IP from DHCP.

    When trying to connect via noVNC (dougw/novnc docker container) there is no other vnc connections to the VM.

    I also have a docker container binhex/arch-krusader which is a file manager connecting via noVNC working good.

    In Firefox Web Developer Tool, the Network traffic shows 2 errors,

    1. Firefox can’t establish a connection to the server at wss://myomvservername.local:8081/websockify.

    2. Failed when connecting: Failed to connect to server ( (code: 1006)) (very generic err code, not much joy googling that)

    Is there anything else I could check or info I can supply to track down my issue?

    Yes. Use qemu-img convert to convert the virtualbox disk to qcow2. Create a new VM and select the converted disk. Maybe I will look at adding a utility to the plugin to help with this.

    Hi ryecoaaron, just tried the convert from virtualbox to qcow2 and got the following error.

    Does it need a Save To new name option?

    Hi All, same problem here, and "usershare path=" in the SMB/CIFS Advanced settings solved the Syslog entry errors.

    The errors in the Syslog were all from old shares from a previous OMV installation, how does that happen?


    My question is, How do old samba shares end up in a fresh install of OMV?


    History of OMV installation and hardware specs may shed some light.


    CPU Xeon E3-1225 V2, 16Gb RAM

    2 x PCIe SATA3.0 6-Port Expansion Controllers

    1 x Dual 1GB NIC

    10 x various sized HDD's configured via MergerFsFolders for NAS shares

    1 x OS drive SSD 1TB


    1st OMV install, March 2020

    OS installed on SSD 1TB

    10 x various sized HDD's configured via MergerFsFolders for NAS shares

    Docker, Portainer installed, multiple containers running, Plex, AirSonic etc

    All worked well until mid March 2021 when I started trialing the new Beta KVM plugin (I'm not blaming the plugin, I screwed something else up)

    Was able to create new Linux VM's and run them

    Tried to create a Windows based VM, but could not.

    Tried to create another Linux based VM, but could not.

    I had screwed something up somewhere, but couldn't recall what I had changed.


    Thought to myself, time for a fresh install of OMV (since it's had 12 months of me pocking n pulling it apart n wrecking it)


    2nd OMV install, mid March 2021

    Physically disconnect all HDD's in MergerFsFolders set up

    Physically disconnect old OS drive 1TB

    Added new SSD 500GB for OS drive

    Installed OMV 5.5.11 onto 500GB SSD, installed any available updates for OMV to version 5.6.2-1, added OMV-Extras, all good so far.

    Physically reconnected 10 HDD's and 1 SSD 1TB (with old installation in place on it).

    Reconfigured OMV and SMB/CIFS shares, but differently to the 1st OMV installation (always looking to improve my set up)

    Installed Docker (installed to 2nd SSD (old OS drive) /srv/d02/docker), Portainer, Yacht and Cockpit.

    All shares accessible via LAN connected Windows, Linux PC's and TV's (Plex working good)

    Installed KVM plugin

    Recreated VM based on previous (working) LinuxOs.qcow2 file. I wouldn't start. Hmmm.

    Created a new linux VM, won't start....... Back to step 1.


    But that's all an aside from How do old samba shares end up in a fresh install of OMV?

    Some how OMV, or more to the point Samba has picked the old OMV installation records.

    How does that occure?

    A hint may lay in the Syslog error below. (### denotes old OMV share name)

    process_usershare_file: stat of /var/lib/samba/usershares/### failed. No such file or directory


    PS Sorry about the essay length post.

    Read-only file system

    Figured out my "Failed to connect to socket: No such file or directory" error

    Directly related to "Failed to bind socket: Read-only file system"

    OS drive got flagged as Read-only file system because "guess who" manually edited the /etc/fstab and /etc/openmediavault/config.xml files.

    Doh, won't do that again.

    Thought I'd be smart and remove the references to a failed HD, the thing is I use mergerfs, that then makes 3 references to the HD in each file, I got 5 out 6 references removed, doh.

    So that was my saturday, figuring out getting my OMV OS drive to be read n write again.

    OMV is very smart, it finds a problem, lock the os drive to stop any corruption.


    Solution was to connect the drive to another linux os, mount, edit fstab n config.xml (double, triple, quadruple check they r the same), disconnect, reboot, OMV is happy again.


    And me, very relieved, n big lessons learnt.

    But then again, I learnt what I shouldn't do and what I can actually do.

    Oh well all in the name of learning.

    That's it for me now with replying to myself.

    Back to playing with KVM plugin...

    ryecoaaron after a bit of advice. I had a drive fail today, see following quote.


    Would a fresh install of OMV5 be a good idea as I'm having network trouble running noVNC ?

    And if that is the case how can I get the VM.qcow2 off a read only file system?

    Or are these questions to much for this thread?