On 22 May 2014, at 08:48, Markus Stockhausen wrote:
Hello,
our hypervisor nodes are running on FC19. virt-preview repos is activated. Windows VMs are using SPICE protocol.
From time to time we see hangups of the windows machines. QEMU has 100% CPU and a perf report gives the following output:
Samples: 70K of event 'cycles', Event count (approx.): 49527366887
51,76% qemu-system-x86 libspice-server.so.1.8.0 [.] 0x000000000005a6e8 41,15% qemu-system-x86 [vdso] [.] 0x0000000000000701 2,09% qemu-system-x86 libc-2.17.so [.] clock_gettime 1,10% qemu-system-x86 libspice-server.so.1.8.0 [.] pthread_self@plt 0,60% qemu-system-x86 libpthread-2.17.so [.] pthread_self 0,53% qemu-system-x86 libspice-server.so.1.8.0 [.] clock_gettime@plt 0,27% qemu-system-x86 qemu-system-x86_64 [.] 0x00000000001d6238 ...
No SPICE client is connected to the VM. Any idea how to isolate the error better?
disable spice i suppose resolve this how about qxl drivers? installed or not? definitely check on spice and/or qemu list… Thanks, michal
Markus <InterScan_Disclaimer.txt>_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users