@ceh-u It’s certainly a curios one as, I’m assuming the exact same image, booted here with full internet access on my virt-manager based KVM setup.
Out of curiosity, given the actual failure message within the VM was name resolution, and we have jumped this ‘no network’ within the KM, you could try pinging, from within the VM, just an internet number i.e one of Googles DNS services.
localhost:~ # ping -c3 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=113 time=12.0 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=113 time=13.1 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=113 time=15.0 ms
--- 8.8.8.8 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2002ms
rtt min/avg/max/mdev = 12.039/13.414/15.055/1.249 ms
As if that works, as above, then there is a network connection but a failing name resolution. I’ve only seen this in earlier Leap i.e. 15.0 15.1 but this is a 15.2 RC image. But that would still narrow stuff down and if it is just name resolution, i.e. name to number resolver related, then you could look to if that is a know issue for KVM guests in your particular version of Ubuntu. As the default network, in this case, is a virtual one provided by the host OS where there is a virtual ‘router’ that forwards, or is expected to forward, such things as DNS requests. And if that is failing then you have narrowed down the area of difficulty. And are likely to learn some network stuff along the way.
Hope that helps and we should have a fully fledged installer out pretty soon anyway. But it will still require full internet access so this situation will still be an issue, at least for your KVM setup if that is the root of the cause. And again, the version of your Ubuntu install will likely help others here help you with these generic network issues, especially if they do relate to the host setup of the virtual network.
Just a thought as there may be some linux networking interest along the way. Or not : one has to pick ones battles; if possible.
Hi Phil – no luck, I’m afraid. I made sure the host network was working, and tried from the guest machine with both Virtio and e1000e as the Device model for the NIC. Both returned “Network is unreachable”.
I have one more thing to try. It involves trying a VM manager on my macOS native system.
Wow! Success. I installed VirtualBox on my Mac, running macOS Catalina 10.15.5, and first time I downloaded the same kvm and xen image. All well except that, as in Ubuntu, at first boot in VB it says it can’t find a boot drive. So I checked a guide on installing in VB and they used the Vmware image (.vmdk). Worked like a charm. And the network is working fine.
So I think that probably Ubuntu on my hardware isn’t quite tuned in and that caused some problems. Thinking back, it did stumble with my Wi-Fi network sometimes. I wonder how VB would go in Ubuntu using the .vmdk image? I’ll try that tomorrow and also install nano.
I could never have got this far without all of your help. This is definitely the best NAS user forum I’ve seen, and I’ve surfed them all! Sincere and enthusiastic thanks.
And yes it does very much look like a network config issue on the Ubuntu / KVM side, hence the enquiry re how old the Ubuntu install is. Ubuntu are normally pretty good however on default configs. Virtual box I believe default to using a bridge arrangement with the host network which is often easier to get right but also means the guest machine is visible to other machines on the LAN. Not always what one wants.
Yes qcow2 is the default in KVM and vmdk in Virtual Box.
That would be good to know, and yes you would have to use the vmdk image given it’s native to the Virtual box system.
Well done getting this sorted. It really should have worked first try. But as you say likely cause is back in history somewhere. But still, these hick-ups can lend themselves as opportunities. But can also be super frustrating.
Given you now have a generic Leap 15.2 RC installed, all be it from a pre-installed disk image, you might want to start a new thread focused on the next issue you run into. It might make for easier searching for folks coming across the same issues.
Chuffed your now out of the woods on this silly network thing, and a plus of options I guess, ie KVM / Virtual Box in this case. I still favour KVM myself but Virtual Box is very popular so that can help on the getting help side. What ever works for you really and given this is only a stepping stone you can at least have a play around now which is nice. Just be sure the drives have a serial assigned to them. In KVM / virt manager they default to this with sata but need it added in virtio. Rockstor depends on these to work properly.
Glad you like the forum, and do feel free to flag anything that seems untoward as we depend on the community to keep things healthy. We do have occasional spam so just flag what you see.
I forgot you asked about the age of Ubuntu - it’s 20.04 LTS, so the most recent stable version. Unfortunately there is no Mac version of VM Manager, so I’ll try it in Ubuntu on any new step, but VB certainly was easy using the vmdk image - hadn’t realised it was native to them. I do think my hardware build may have been a bit of a challenge for Ubuntu, e.g. the Wi-Fi device is something that works well with Apple but may not be common otherwise.
I will certainly now switch to new posts for any following stuff. This one was certainly frustrating on one hand but hugely encouraging for the support you’ve given.
I’m glad you’re now sorted, but I was curious about one thing on the screenshot you shared: Is that really all you have printed out on the screen when you turn on the JeOS_Leap15.2RC VM (using Virtual Machine Manager)?
I personally see the normal system logs at boot:
Do you mean was other information also listed? You are probably referring to the last screenshot. The usual log information is exacly as you have shown. For whatever reason, that last output was exactly as shown in the screenshot. It was following a reboot and I remember wondering why; I thought I may have only paused it but the first line about the failure to check a link status was always the first line or so of the log.
I was, indeed, referring to your last screenshot… sorry I wasn’t clear; if you look at the top right corner of my post and click on the “reply arrow” icon next to your name, you’ll see the post I replied to but it’s not the most intuitive thing and I should have been more explicit.
OK… that’s what I was wondering. It does seem the problem lies in the network configuration (host, or through Virtual Machine Manager) in this case. I do see the same Network is unreachable when the host is using a different connection than usual (a corporate VPN, for instance), so there’s probably some network misconfiguration there, for instance. As @phillxnet mentioned, it’s probably not worth your time troubleshooting that one for now given you have a running solution using VirtualBox and that a new Rockstor installer will arrive soon. In case you’re interested, though, you can have a look at openSUSE documentation on networking. Their documentation can be very informative and detailed… see below, for instance: https://doc.opensuse.org/documentation/leap/reference/html/book.opensuse.reference/cha-network.html
Hi Flox - thank you for the input. I believe Ubuntu is not quite comfortable with my system’s hardware, so that affected the VM manager and guest. But VirtualBox has handled it with no trouble, which probably reflects the disparity in the reources of Oracle and VM Manager. Anyway, I can forget about that now. VirtualBox does the job both on macOS and Ubuntu, so I’ll work in Ubuntu.
But for interest I will review the documentation you referred to.
Thanks again.
I decided to try Virtual Machine Manager in a full Leap 15.2RC install. Reason being I couldn’t get the VM guest JeOS to connect to the network in Ubuntu, and I thought it possible that might work better in a full-bore openSUSE install. Also, in VirtualBox I couldn’t find any way to add a serial number to the virtual disk as required by Rockstor.
I had to install a lot of VM-related stuff in Leap for VMM. It connected to the network fine, and I now have the virtual disk serial number added in the JeOS guest. The full Leap 15.2 is a very impressive system.
@ceh-u So could you clarify your current setup. I.e. the host OS and the guest OS.
So am I right in thinking that you now have Leap 15.2 RC as your host OS, the one running on bare metal. And you are using Virtual Machine Manager into which you have installed a Leap 15.2 RC JeOS pre-installed image (as before)? And can you ping by name an internet location from within the JeOS pre-installed image, so you can install nano to help with what’s next?
And out of interest if the above is correct, and unrelated to this discussion, which desktop did you chose?
I know it’s been a bit confusing. I’m lucky to have 3 SSDs: one running macOS, one that has Ubuntu 20.04, and now one that has Leap 15.2 RC, each running on the bare metal. On the Leap system I have VMM with the Leap 15.2RC JeOS preinstalled image, and from that I can ping by name Internet locations. The version of VMM on the Leap desktop seems more sophisticated than the one I tried in Ubuntu. When I first installed it through Software after a search and opened it, there were messages saying various virtlib modules were missing. In Software I found a lot of stuff in YaST2 modules, including VMM, so I asked to install likely looking ones. Anyway, VMM worked well and the JeOS image connected to the Internet straight away.
If you mean which desktop for running VMM, it will be the Leap desktop. I hope that’s clearer.
OK spoke too soon. This morning in the JeOS VM guest, there is an error(?) triangle next to Network source: Virtual network ‘default’:NAT in the Virtual Network interface window. I did see when the guest was booting a series of messages about starting the wicked network, with time limits.
I take it you mean host here. I.e. the OS running on the bare metal. That is the one that sets up the environment within which the guest OS (the one inside the Virtual Machine) uses to get it’s network connection. So likely changes in the host (outer layer) have left the guest (inner layer) out of kilter. This stuff is usually all automatic and given you have observed the same thing across multiple host OS’s it’s looking like something to do with how you have your networking set up. I’d guess that the change from one network type ‘wired’ to another ‘wifi’ is throwing off the virtual inner network configuration as it would need to adapt to the outer (host) changes from wired ethernet to wifi and back and is maybe just failing in that regard. Are you often changing between them. Or maybe your machine does this itself often as maybe one is intermittent. That is a challenging environment to have a virtual network and it’s associated guests to work within. Try sticking to ethernet and rebooting the host so it can re-setup a hopefully appropriate and working, virtual network within which the guest can then boot and get it’s bearings (read internet connection).
All very much outside the realm of Rockstor I’m afraid and I’m personally not in favour of Rockstor virtual instance use as it is, in my eyes, intended for hardware only really. But VMs are perfect for playing around, and development of course.
Hi Phil – The only reason I changed (from wi-fi to wired) was to see if that would suit VMM better. I have used wi-fi for a long time but wired should be a more consistent connection.
What happened today was that, looking at the YaST2 clutch of modules there was one labelled Create Virtual Machine as well as the VMM one - same logo. So I tried it and it had more options for setting up another virtual network (as well as the default). But if trying its settings they didn’t seem to be accepted any better (no IP appearing, etc). So I changed back to the default, using the wired connection again. I got an IP address, even though there was still a little orange triangle alongside the virtual network item. So I rebooted and tried the machine made with that module and it pinged without any trouble, and I was able to install nano in the JeOS guest.
That’s not to say it will work consistently, I’ll just have to hope and see. Anyway, at this stage I’m ready for the next step.
: one has to pick ones battles; if possible.
