> Unable to complete install: 'unsupported configuration: CPU mode 'custom' for x86_64 kvm domain on x86_64 host is not supported by hypervisor'
with virt-manager.
I think I checked something called "custom" somewhere. But then I unchecked it.
I switch to "copy host CPU configuration". I'm now getting:
> Unable to complete install: 'internal error: process exited while connecting to monitor: Could not access KVM kernel module: Permission denied failed to initialize KVM: Permission denied'
@yode You can't ping people in chat unless they have been in the room in the past two weeks. If you want to invite someone to chat, leave a comment under their question or answer and ask them to come here.
Of course we know cat /proc/cpuinfo |grep "cpu cores" will give a output
[root@14:47 ~]# cat /proc/cpuinfo |grep "cpu cores"
cpu cores : 2
cpu cores : 2
cpu cores : 2
cpu cores : 2
But actually I want to get the total number of cpu cores.I mean I hope the result is
cp...
Also, @yode Daniel is essentially asking if you want the number of physical CPU chips (what you would count if you opened your machine) or the number of CPUs those chips have (most today have >1) or the number of cores the machine will see (most CPUs have multiple cores)
SO on my laptop, I have one physical chip (i7-6820HQ) which has 4 physical cores, each of which has 2 logical cores (@StephenKitt please correct me if I'm wrong).
This means that in practical terms, I can run 8 processes in parallel each of which will get a full "CPU" and that is why I get 8 when looking at /proc/cpuinfo:
Of course we know cat /proc/cpuinfo |grep "cpu cores" will give an output
[root@14:47 ~]# cat /proc/cpuinfo |grep "cpu cores"
cpu cores : 2
cpu cores : 2
cpu cores : 2
cpu cores : 2
But actually I want to get the total number of cpu cores. I want the result to be
cpu ...
@FaheemMitha remember virt-manager is just a frontend to libvirt—libvirt is what is actually running the VM. What user you run virt-manager doesn't change that.
libvirt is the server. virt-manager is the client. libvirt is what actually runs the VM; virt-manager is a client that connects to libvirt to give it commands
> Jul 12 05:39:52 orwell libvirtd[20045]: 2017-07-12 00:09:52.408+0000: 20046: error : qemuConnectGetDomainCapabilities:19016 : invalid argument: KVM is not supported by '/usr/bin/kvm' on this host
I don't know what that means, but it looks relevant.
And there's some stuff about CPU, which goes away if one chooses the same as host.
> Jul 12 05:42:18 orwell libvirtd[20045]: 2017-07-12 00:12:18.272+0000: 20047: error : qemuProcessUpdateGuestCPU:5136 : unsupported configuration: CPU mode 'custom' for x86_64 kvm domain on x86_64 host is not supported by hypervisor
But I'll try again, in real time. Sigh.
It's one of those days when it feels less surprising that most people don't run things like Debian.
@FaheemMitha odd... if you run /usr/bin/kvm, what happens? I guess maybe nothing if you're not a member of the kvm group, but here it pops up a window and attampts a netboot
if you care: I think it works as a side-effect of enabling the persistent journal, at least if you're a member of the adm group. Instructions are in /usr/share/doc/systemd/README.Debian.gz
@FaheemMitha sestatus will tell you your current SELinux mode (first status line and current mode, if any)
@ÉtienneBézout if it's moderately involved, you may get more visibility on the main site than in chat; no harm in asking in chat, just much smaller audience
It might be a bit vague. My problem is this: When I click on the network manager icon, it says I have a connection, albeit slow (10 Mbs). However, I cannot ping the router. If I lof into the router, the device does appear.
@ÉtienneBézout It'd be useful to install ethtool if you can (obviously, without a network connection that's harder!) Also, have you tried a different Ethernet cable?
And of course, I should ask, is the router configured to respond to pings? Just to avoid an "oh... DUH" moment later.
Right, but a can't get to the Internet could be all kinds of config issues (e.g., firewall rules, NAT rules, wrong gateway...). There can always be multiple issues, unfortunately.
@ÉtienneBézout If the activity light is flashing randomly (not in a fixed pattern), that's probably a good sign. Is there an activity light on the router end as well?
Also, have you double-checked your IP settings? Though I suspect the link isn't working, if DHCP didn't work.
Next thing to check would be ip link ls dev DEVICE to make sure it shows as both UP and LOWER_UP. Other than that ip addr ls dev DEVICE to make sure Network Manager actually configured the IP address on it.
I just checked and I can ping the router from another computer on the LAN. Yes, there does appear to be some random flashing. There's also some flashing on the switch, which is connected to my router.
If you had ethtool installed, it's possible ethtool -S enp5so would give more details. But it's most likely a hardware issue. I'd try a different cable before binning the card. Given, an Ethernet card is $10...
A marginal cable might work with one card but not another. So worth trying another cable, if you have one around.
Also take a moment to physically inspect the port: make sure, e.g., there isn't a bunch of lint in it. Or a bent pin.
Well, it's not so easy to try a different cable due to the wiring. Also, as I said I tried the cable on other computers (of the same model). One of them got 100 Mb/s and the others got 10 Mb/s. Ok, I'll check the port. Strange though that TX would work but not RX?
Ah, ok maybe it is a cable issue. It was kind of a poor job when installed with. Small radius of curvature and keystone jack on stranded cable :( Still, worked OK for one computer.
@derobert I'd try a reboot first. But let me first ask this. If libvirt was installed, then might it not take a reboot to add libvirt-qemu to the kvm group?
@ÉtienneBézout BTW: If you're paying someone to run in-wall cabling, they #*@@ ought to do the job right with cat5e or cat6 solid copper. And then prove its done right by testing it with e.g., a Fluke cable tester
@FaheemMitha oh, on chat? I think they ought to do something where it's related to how busy the room is. You don't want someone editing a message that's scrolled off screen (or nearly so) where no one will see the edit... 5m can be a lot in an active chat. But less active ones, it'll still be the last or 2nd to last message. So then it'd be OK.
@FaheemMitha That probably means you don't have SELinux at all
@dreobert No, I did it myself. I pulled the cable through an old water pipe, and some turns were a bit sharp. I took a risk with it, but since it worked with one computer I thought it was ok.
@FaheemMitha sure, why not? Only downside is you'll have to close everything you're working on, and then carefully re-set-up 12 desktops worth of xterms... (Can you tell I hate rebooting?)
So, can you show me i) the output of ls | head in the directory with the bad files so I can see a few; ii) the exact command you used to rename them (as in copy/paste from your terminal).
Also, what operating system are you doing this in? And what shell?
(oh, and if you don't have time, don't let me keep you. I pinged you because I had to in order to create your chat account and make sure you could write here. If you're in a rush, don't worry, we can do it another time).
@AndyP btw: once you paste that in, type another space then the 'fixed font' button will appear. Delete the extra space and then tap the button. That'll get your paste formatted nicely in chat.
We are currently testing a new application on Linux that previously worked on Windows. It appears there were some hard coded paths in the program and a bunch of test files with long filenames with backslashes were created. For example:
directory\subdirectory\subdirectory\subdirectory\filename.e...
Of course we know cat /proc/cpuinfo |grep "cpu cores" will give a output
[root@14:47 ~]# cat /proc/cpuinfo |grep "cpu cores"
cpu cores : 2
cpu cores : 2
cpu cores : 2
cpu cores : 2
But actually I want to get the total number of cpu cores.I mean I hope the result is
cp...
