On Thu, Jun 01, 2017 at 08:47:55AM -0400, Michael C. Cambria wrote:
Hi,
Ever since I recently upgraded to Fedora 25, I can't get kvm working.
It's not really about KVM but about libvirt. So the daemon is running...
It's worked on this system since initial fedora 20 install. All
upgrades were done via yum, then once available, dnf.
I do have libvirt-sock in LISTENING state: STREAM LISTENING
/var/run/libvirt/libvirt-sock
I noticed I also have multiple connections in CONNECTING state:
STREAM CONNECTING 0 /var/run/libvirt/libvirt-sock
STREAM CONNECTING 0 /var/run/libvirt/libvirt-sock
STREAM CONNECTING 0 /var/run/libvirt/libvirt-sock
I rebooted and tried again. Each e.g. 'virsh -c" or virt-manager
command I issue, I end up with another socket in CONNECTING state.
restarting libvirtd.service close them leaving just one LISTENING
strace for virsh -c shows:
4301 socket(AF_UNIX, SOCK_STREAM, 0) = 5
4301 connect(5, {sa_family=AF_UNIX,
sun_path="/var/run/libvirt/libvirt-sock"}, 110) = 0
4301 getsockname(5, {sa_family=AF_UNIX}, [128->2]) = 0
4301 futex(0x7f3c0a2e4fc8, FUTEX_WAKE_PRIVATE, 2147483647) = 0
4301 fcntl(5, F_GETFD) = 0
4301 fcntl(5, F_SETFD, FD_CLOEXEC) = 0
4301 fcntl(5, F_GETFL) = 0x2 (flags O_RDWR)
4301 fcntl(5, F_SETFL, O_RDWR|O_NONBLOCK) = 0
virsh is just waiting for the daemon, there's no useful info.
systemctl status libvirtd.service and journalctl -f -u
libvirtd.service
do not show any log entries for libvird after it has been started. e.g.
it doesn't log anything about the connection attempts.
Anyone have an idea where to look next?
Check if the daemon is running, check any possible QEMU processes that
might be running, enable debug logs for libvirtd [1] and try grabbing a
backtrace of the process (For example `gdb -batch -p $(pidof libvirtd)
-ex "t a a bt full"' after you install debuginfo packages).
Thanks,
MikeC
_______________________________________________
libvirt-users mailing list
libvirt-users(a)redhat.com
https://www.redhat.com/mailman/listinfo/libvirt-users