Description of problem: When adding new hardware to an exists guest, virt-manager will suspend. Version-Release number of selected component (if applicable): Host Machine: RHEL5.4 Server X86_64 20090701.0 virt-manager-0.6.1-5.el5 libvirt-0.6.3-13.el5 How reproducible: 100% Steps to Reproduce: 1.Start libvirtd service, run virt-manager. 2.Choose an exists guest and open the virtual machine window, try to add some new hardware (Click "Hardware" Tab, and then click "Add Hardware"). Actual results: When "Add Hardware" button clicked, the virt-manager will have none response, check the process by top command, the libvirtd service take 100% CPU usage. The "Adding new Hardware" window will popup after restart libvirtd service. Close virt-manager and reopen it, and try to add some new hardware again, this issue occurs again too. Expected results: The virt-manager should can add hardware successfully. Additional info:
Well, I am not 100% sure but doesn't this seem to be libvirtd issue? According to those information I guess libvirtd is "stuck in the loop". The libvirtd service 100% CPU usage is the sign of this (I guess). In fact I never found this issue, could you provide more information about that? I have tried it this way - open virt-manager, open a VM window and selecting Hardware tab and clicking Add hardware. As a device I selected "Sound" and type "es1370", clicked Next and then clicked Finish. Everything went fine so what exactly did you tried to add? Wasn't is some physical host device that was not responding so libvirtd got stuck in the loop or something?
Hi Michal, I can also reproduce this issue in my machine that I installed just now. Version-release: RHEL5.4-Server-x86_64(2.6.18-156.el5xen) # rpm -qa | grep -E "libvirt |virt-manager|xen" virt-manager-0.6.1-5.el5 xen-3.0.3-88.el5 kernel-xen-2.6.18-156.el5 xen-libs-3.0.3-88.el5 xen-libs-3.0.3-88.el5 kernel-xen-devel-2.6.18-156.el5
I am having the same version of xen, kernel-xen and xen* stuff and I added it as an audio device. It was working fine. Also, sorry for confusion when writing about Physical Host Device. I saw this in virt-manager only for KVM, not Xen so sorry for confusion. Anyway, I was unable to reproduce it at all so that's strange. Can't this be a hardware related issue? What version (i386/x86_64/ia64) and what hardware do you use for that? I tend to say this is something libvirtd related rather than virt-manager according to 100% CPU usage of libvirtd service.
Also, libvirtd is the same as in c#0 - 0.6.3-13.el5 so everything seems to be the same.
(In reply to comment #2) > Hi Michal, I can also reproduce this issue in my machine that I installed just > now. > > Version-release: > RHEL5.4-Server-x86_64(2.6.18-156.el5xen) > > # rpm -qa | grep -E "libvirt |virt-manager|xen" > virt-manager-0.6.1-5.el5 > xen-3.0.3-88.el5 > kernel-xen-2.6.18-156.el5 > xen-libs-3.0.3-88.el5 > xen-libs-3.0.3-88.el5 > kernel-xen-devel-2.6.18-156.el5 My hardware information and kernel information: #uname -a Linux dhcp-66-70-37.nay.redhat.com 2.6.18-156.el5xen #1 SMP Mon Jun 29 18:24:43 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux And OS version is rhel5u4 20090701.0 (beta) #cat /etc/redhat-release Red Hat Enterprise Linux Server release 5.4 Beta (Tikanga) Guest info: # cat /etc/xen/pv-rhel name = "pv-rhel" uuid = "722a19b2-ff5d-3125-80e5-114bf821d3e0" maxmem = 2048 memory = 512 vcpus = 4 bootloader = "/usr/bin/pygrub" on_poweroff = "destroy" on_reboot = "restart" on_crash = "restart" vfb = [ "type=vnc,vncunused=1,keymap=en-us" ] disk = [ "tap:aio:/var/lib/xen/images/pv-rhel.img,xvda,w" ] vif = [ "mac=00:16:36:49:ff:87,bridge=virbr0,script=vif-bridge" ] Host CPU info: #cat /proc/cpuinfo processor : 3 vendor_id : GenuineIntel cpu family : 6 model : 23 model name : Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz stepping : 10 cpu MHz : 2826.312 cache size : 6144 KB physical id : 3 siblings : 1 core id : 0 cpu cores : 1 fpu : yes fpu_exception : yes cpuid level : 13 wp : yes flags : fpu tsc msr pae mce cx8 apic mtrr mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm constant_tsc pni monitor ds_cpl vmx smx est tm2 cx16 xtpr lahf_lm bogomips : 7068.81 clflush size : 64 cache_alignment : 64 address sizes : 36 bits physical, 48 bits virtual power management:
Additional, we've tried a RHEL5U3-Server-X86_64, and update the related packages to the above version, the "Add hardware" feature works successfully, the same phenomenon as your describe. But the RHEL5U4-Server-x86_64 (20090701.0) do not work on all 3 kinds of guest( KVM, Xen FV, Xen PV), it's 100% reproducable. Does that means some other packages related to this issue?
Ok, could you grab ~/.virt-manager/virt-manager.log file from this test and append it to this BZ ?
Created attachment 350762 [details] ~/.virt-manager/virt-manager.log
There are lots of errors in the log. But what concerns me is the broken pipe "[Wed, 08 Jul 2009 01:31:57 virt-manager 6531] DEBUG (error:76) Uncaught Error: Uncaught error validating hardware input: cannot send data: Broken pipe". I don't know whether it can be Xen related or not. Could you please try using *test* RPMs with Intel patches ? You can download them from http://people.redhat.com/minovotn/xen . Please let me know the results with those *test* packages and version of components (virt-manager/libvirtd) described above (c#2). Thanks, Michal
The reason 'Add Hardware' hangs is because of libvirt bug 509500 : when launching the wizard, we are doing the equivalent of 'virsh nodedev-list', which causes libvirtd to get stuck.
Cole, then I guess this is a dup of 509500, right? Or at least libvirtd issue and not virt-manager one.
closed as a duplicate. *** This bug has been marked as a duplicate of bug 509500 ***