- Users - libvirt List Archives

[libvirt-users] virsh list show shutdown VMs only
by Heiko L. 23 Mar '13

23 Mar '13

Hallo, I am using virsh with squeeze and xen-4.2+linux-3.8 (compiled) Problem: "virsh list -all" shows VMs only in shutdown state. (see: test3 virsh list shutdown) virsh start/shutdown works correctly. I suspect that version of libvirt is to old. Is their some workaround without updating system? regards Heiko details: - test1 start vm root@e2:/tmp# virsh start vm1_test1 Domain vm1_test1 started root@e2:/tmp# xm list Name ID Mem VCPUs State Time(s) Domain-0 0 64487 8 r----- 6834.7 vm1_test1 6 96 1 -b---- 2.3 root@e2:/tmp# virsh list Id Name State ---------------------------------- root@e2:/tmp# virsh list --all Id Name State ---------------------------------- root@e2:/tmp# -> vm running, but "virsh list" show nothing - test2 debug virsh list root@e2:/tmp# LIBVIRT_DEBUG=1 virsh list | egrep -v "do_open|virRegister" 14:36:38.540: debug : virInitialize:340 : register drivers 14:36:38.541: debug : xenHypervisorInit:2010 : Using new hypervisor call: 40002 14:36:38.542: debug : xenHypervisorInit:2123 : Failed to find any Xen hypervisor method 14:36:38.542: debug : vboxRegister:122 : VBoxCGlueInit failed, using dummy driver 14:36:38.543: debug : virConnectOpenAuth:1499 : name=(null), auth=0x7f46145b2700, flags=0 14:36:38.543: debug : xenUnifiedOpen:326 : Trying hypervisor sub-driver 14:36:38.543: debug : xenUnifiedOpen:329 : Activated hypervisor sub-driver 14:36:38.543: debug : xenUnifiedOpen:337 : Trying XenD sub-driver 14:36:38.546: debug : xenUnifiedOpen:340 : Activated XenD sub-driver 14:36:38.546: debug : xenUnifiedOpen:353 : Trying XS sub-driver 14:36:38.547: debug : xenStoreOpen:345 : Failed to add event handle, disabling events 14:36:38.547: debug : xenUnifiedOpen:356 : Activated XS sub-driver 14:36:38.551: debug : xenUnifiedOpen:392 : Trying Xen inotify sub-driver 14:36:38.551: debug : xenInotifyXendDomainsDirLookup:139 : Looking for dom with uuid: d868acf8-56b7-b260-4c85-c2e1e2a67e9b 14:36:38.576: debug : virGetDomain:382 : New hash entry 0x65b0a0 14:36:38.576: debug : virDomainFree:2243 : domain=0x65b0a0 14:36:38.576: debug : virUnrefDomain:463 : unref domain 0x65b0a0 vm1_test1 1 14:36:38.576: debug : virReleaseDomain:416 : release domain 0x65b0a0 vm1_test1 d868acf8-56b7-b260-4c85-c2e1e2a67e9b 14:36:38.576: debug : virReleaseDomain:433 : unref connection 0x634ad0 2 14:36:38.576: debug : xenInotifyOpen:441 : Adding a watch on /var/lib/xend/domains 14:36:38.576: debug : xenInotifyOpen:453 : Building initial config cache 14:36:38.576: debug : xenInotifyOpen:460 : Registering with event loop 14:36:38.576: debug : xenInotifyOpen:464 : Failed to add inotify handle, disabling events 14:36:38.576: debug : xenUnifiedOpen:395 : Activated Xen inotify sub-driver 14:36:38.576: debug : doRemoteOpen:565 : proceeding with name = xen:/// 14:36:38.576: debug : remoteIO:9888 : Do proc=66 serial=0 length=28 wait=(nil) 14:36:38.576: debug : remoteIO:9963 : We have the buck 66 0x7f461474d010 0x7f461474d010 14:36:38.577: debug : remoteIODecodeMessageLength:9316 : Got length, now need 64 total (60 more) 14:36:38.577: debug : remoteIOEventLoop:9814 : Giving up the buck 66 0x7f461474d010 (nil) 14:36:38.577: debug : remoteIO:9992 : All done with our call 66 (nil) 0x7f461474d010 14:36:38.577: debug : remoteIO:9888 : Do proc=1 serial=1 length=48 wait=(nil) 14:36:38.577: debug : remoteIO:9963 : We have the buck 1 0x65e2d0 0x65e2d0 14:36:38.608: debug : remoteIODecodeMessageLength:9316 : Got length, now need 56 total (52 more) 14:36:38.608: debug : remoteIOEventLoop:9814 : Giving up the buck 1 0x65e2d0 (nil) 14:36:38.608: debug : remoteIO:9992 : All done with our call 1 (nil) 0x65e2d0 14:36:38.608: debug : doRemoteOpen:942 : Adding Handler for remote events 14:36:38.608: debug : doRemoteOpen:949 : virEventAddHandle failed: No addHandleImpl defined. continuing without events. 14:36:38.609: debug : doRemoteOpen:565 : proceeding with name = xen:/// 14:36:38.609: debug : remoteIO:9888 : Do proc=66 serial=0 length=28 wait=(nil) 14:36:38.609: debug : remoteIO:9963 : We have the buck 66 0x69e3b0 0x69e3b0 14:36:38.609: debug : remoteIODecodeMessageLength:9316 : Got length, now need 64 total (60 more) 14:36:38.609: debug : remoteIOEventLoop:9814 : Giving up the buck 66 0x69e3b0 (nil) 14:36:38.609: debug : remoteIO:9992 : All done with our call 66 (nil) 0x69e3b0 14:36:38.610: debug : remoteIO:9888 : Do proc=1 serial=1 length=48 wait=(nil) 14:36:38.610: debug : remoteIO:9963 : We have the buck 1 0x69e3b0 0x69e3b0 14:36:38.643: debug : remoteIODecodeMessageLength:9316 : Got length, now need 56 total (52 more) 14:36:38.643: debug : remoteIOEventLoop:9814 : Giving up the buck 1 0x69e3b0 (nil) 14:36:38.643: debug : remoteIO:9992 : All done with our call 1 (nil) 0x69e3b0 14:36:38.643: debug : doRemoteOpen:942 : Adding Handler for remote events 14:36:38.643: debug : doRemoteOpen:949 : virEventAddHandle failed: No addHandleImpl defined. continuing without events. 14:36:38.643: debug : doRemoteOpen:565 : proceeding with name = xen:/// 14:36:38.643: debug : remoteIO:9888 : Do proc=66 serial=0 length=28 wait=(nil) 14:36:38.643: debug : remoteIO:9963 : We have the buck 66 0x6de490 0x6de490 14:36:38.645: debug : remoteIO:9888 : Do proc=1 serial=1 length=48 wait=(nil) 14:36:38.645: debug : remoteIO:9963 : We have the buck 1 0x6de490 0x6de490 14:36:38.677: debug : remoteIODecodeMessageLength:9316 : Got length, now need 56 total (52 more) 14:36:38.677: debug : remoteIOEventLoop:9814 : Giving up the buck 1 0x6de490 (nil) 14:36:38.677: debug : remoteIO:9992 : All done with our call 1 (nil) 0x6de490 14:36:38.677: debug : doRemoteOpen:942 : Adding Handler for remote events 14:36:38.677: debug : doRemoteOpen:949 : virEventAddHandle failed: No addHandleImpl defined. continuing without events. 14:36:38.677: debug : virConnectNumOfDomains:1894 : conn=0x634ad0 Id Name State ---------------------------------- 14:36:38.678: debug : virConnectClose:1525 : conn=0x634ad0 14:36:38.678: debug : virUnrefConnect:294 : unref connection 0x634ad0 1 14:36:38.678: debug : remoteIO:9888 : Do proc=2 serial=2 length=28 wait=(nil) 14:36:38.678: debug : remoteIO:9963 : We have the buck 2 0x6de490 0x6de490 14:36:38.678: debug : remoteIODecodeMessageLength:9316 : Got length, now need 56 total (52 more) 14:36:38.678: debug : remoteIOEventLoop:9814 : Giving up the buck 2 0x6de490 (nil) 14:36:38.678: debug : remoteIO:9992 : All done with our call 2 (nil) 0x6de490 14:36:38.678: debug : remoteIO:9888 : Do proc=2 serial=2 length=28 wait=(nil) 14:36:38.678: debug : remoteIO:9963 : We have the buck 2 0x65e2d0 0x65e2d0 14:36:38.679: debug : remoteIODecodeMessageLength:9316 : Got length, now need 56 total (52 more) 14:36:38.679: debug : remoteIOEventLoop:9814 : Giving up the buck 2 0x65e2d0 (nil) 14:36:38.679: debug : remoteIO:9992 : All done with our call 2 (nil) 0x65e2d0 14:36:38.679: debug : remoteIO:9888 : Do proc=2 serial=2 length=28 wait=(nil) 14:36:38.679: debug : remoteIO:9963 : We have the buck 2 0x65e2d0 0x65e2d0 14:36:38.680: debug : remoteIODecodeMessageLength:9316 : Got length, now need 56 total (52 more) 14:36:38.680: debug : remoteIOEventLoop:9814 : Giving up the buck 2 0x65e2d0 (nil) 14:36:38.680: debug : remoteIO:9992 : All done with our call 2 (nil) 0x65e2d0 14:36:38.682: debug : virReleaseConnect:249 : release connection 0x634ad0 -> entry "unref domain 0x65b0a0 vm1_test1 1" found but not displayed - test3 virsh list shutdown # virsh shutdown vm1_test1 Domain vm1_test1 is being shutdown root@e2:/tmp# virsh list Id Name State ---------------------------------- root@e2:/tmp# virsh list --all Id Name State ---------------------------------- - vm1_test1 shut off -> "virsh list" show vm if shutdown only - version # apt-cache policy libvirt0 libvirt0: Installed: 0.8.3-5+squeeze2 Candidate: 0.8.3-5+squeeze2 Version table: *** 0.8.3-5+squeeze2 0 # uname -a Linux e2 3.8.0-xen0 #1 SMP Fri Mar 1 14:01:33 CET 2013 x86_64 GNU/Linux # xm info | grep ^xen_m xen_major : 4 xen_minor : 2

1 1

[libvirt-users] howto make libvirtd xen-4.2
by Heiko L. 23 Mar '13

23 Mar '13

Hallo I wanted to compile libvirtd on xen-4.2 but on [3] wrote: ...current version of libvirt is not compatible with xen 4.2... Which version that I can test? regards heiko -------------------------------------------------- - details 130305.xen.make.libvirtd #### var tmp=/tmp libvirtver="0.10.2" pkgs="libxml2-dev libdevmapper-dev" #### dl cd $tmp ## dl [2] wget ftp://libvirt.org/libvirt/libvirt-$libvirtver.tar.gz apt-get install $pkgs #### make tar -xzf libvirt-$libvirtver.tar.gz cd libvirt-$libvirtver ./configure make && make install ----------------------------------------------------- #### details # make ... CC libvirt_driver_xen_impl_la-xen_inotify.lo CCLD libvirt_driver_xen_impl.la CCLD libvirt_driver_xen.la CC libvirt_driver_libxl_impl_la-libxl_conf.lo In file included from libxl/libxl_conf.c:43: libxl/libxl_conf.h:61: error: field 'ctx' has incomplete type libxl/libxl_conf.h:80: error: field 'ctx' has incomplete type libxl/libxl_conf.h:81: error: expected specifier-qualifier-list before 'libxl_waiter' libxl/libxl_conf.c: In function 'libxlMakeDomCreateInfo': libxl/libxl_conf.c:365: warning: implicit declaration of function 'libxl_init_create_info' [-Wimplicit-function-declaration] libxl/libxl_conf.c:365: warning: nested extern declaration of 'libxl_init_create_info' [-Wnested-externs] - on [3] wrote: ...current version of libvirt is not compatible with xen 4.2... ========================================================================= [1] https://www.redhat.com/archives/libvirt-users/2012-September/msg00113.html Re: [libvirt-users] virsh list not working with xen 4 virsh list --all only shows turned off machines registered in xend. [2] ftp://libvirt.org/libvirt/ Index of ftp://libvirt.org/libvirt/ [3] http://superuser.com/questions/481662/xen-4-2-on-centos-6-3-cant-compile-a-… Xen 4.2 on CentOs 6.3 : can't compile a libvirt 0.9.10 xen-activated? ...current version of libvirt is not compatible with xen 4.2...

1 1

[libvirt-users] Errors while using blkiotune command
by Saurabh Deochake 23 Mar '13

23 Mar '13

Hi all, I want to limit the I/O bandwidth inside the container, so I used virsh command blkiotune. But when I enter a command: virsh # blkiotune lxcguest --weight 250 I get following errors: error: Unable to change blkio parameters error: Requested operation is not valid: blkio cgroup isn't mounted I also have blkio cgroup mounted. What can be the problem? Thanks in advance. Regards, Saurabh Deochake

4 6

[libvirt-users] [bump] Storage pool created on the "wrong" vg
by Alaric Haag 22 Mar '13

22 Mar '13

________________________________ From: Alaric Haag Sent: Tuesday, March 19, 2013 2:39 PM To: libvirt-users(a)redhat.com Subject: Storage pool created on the "wrong" vg Hello all, (Hopefully, I am posing this question to the correct list!) I seem to have mis-clicked through the creation of an LVM-based storage pool in virt-manager and it is using a volume group containing LVs dedicated to the root/swap filesystems. As these are active LVs, there seems no way to remove this pool. Is that true? My reading thus far suggests that, with LVM storage units, you can't delete them from the pool without REALLY deleting them from the VG. I'm "new enough" to KVM to be missing something obvious, but I'd be pretty shocked if a "structure" like a storage pool can't be deconstructed. Kind regards, Alaric ------------------------ Here's a bit more detail in the hope someone might help: I clicked through the creation of a storage pool, and accidentally used the wrong volume group, so I have: virsh # vol-list vm_pool Name Path ----------------------------------------- lv_root /dev/vg_01/lv_root lv_swap /dev/vg_01/lv_swap These two LVs, in group vg_01, are the / and swap partition of the host system. I meant for the pool to reside on vg_vm, a different volume group. Can I safely run: virsh # vol-delete lv_root --pool=vm_pool virsh # vol-delete lv_swap --pool=vm_pool virsh # pool-delete vm_pool or do the first two steps actually destroy the LVs in vg_01 thus bringing my system crashing down? The link below strongly suggests I cannot delete these logical volumes from the pool, but is a "draft" for OpenSUSE. (I'm running libvirt 0.10.2-18 on RHEL6) http://doc.opensuse.org/products/draft/SLES/SLES-kvm_sd_draft/cha.libvirt.s… If that's the case, is there ANY way to undo my mistake and remove this pool? I am most grateful for any guidance! Alaric

1 0

Re: [libvirt-users] remote connection issue 'virsh -c qemu+ssh:///root@localhost/system list'
by Javi Legido 21 Mar '13

21 Mar '13

Hi Olivia. Another think that you should check is socat package. In my documented example the hypervisor running KVM is a Debian squeeze, and I had to proceed this way: sudo aptitude update; sudo aptitude install netcat socat -R sudo vim /bin/netcatsocket #!/bin/bash socat - unix-client:$2 sudo chmod +x /bin/netcatsocket You can try it and see if it helps or not. PD: looping in again the list, I replied just to Olivia before Cheers El 21/03/2013 12:47, "Yin Olivia-R63875" <r63875(a)freescale.com> escribió: > Hi Javier, > > Thanks for your help. > From the pages you provided, I guess you suggest is as below: > * Cannot recv data: Host key verification failed. : Connection reset by > peer > > Sucede al lanzar: > > virt-manager -c qemu+ssh://usuario@hipervisor/system > > Solucion: iniciar una sesion SSH para almacenar las claves del hipervisor > en el cliente: > > ssh usuario@hipervisor > > Exactly, I've login with ssh and then connect with virsh. > > user@x86:~$ ssh root(a)10.193.20.109 > root(a)10.193.20.109's password: > root@ppc:~# exit > logout > Connection to 10.193.20.109 closed. > > user@x86:~$ virsh -c qemu+ssh://root@10.193.20.109/system list --all > root(a)10.193.20.109's password: > error: failed to connect to the hypervisor > error: End of file while reading data: nc: invalid option -- 'U' > BusyBox v1.19.4 (2013-03-08 13:08:18 CST) multi-call binary. > > Usage: nc [-iN] [-wN] [-l] [-p PORT] [-f FILE|IPADDR PORT] [-e PROG]: > Input/output error > > It still failed to connect the hypervisor, instead of ssh issue. > > Best Regards, > Olivia > > > -----Original Message----- > > From: javilegido(a)gmail.com [mailto:javilegido@gmail.com] On Behalf Of > Javi > > Legido > > Sent: Thursday, March 21, 2013 7:35 PM > > To: Yin Olivia-R63875 > > Subject: Re: [libvirt-users] remote connection issue 'virsh -c > > qemu+ssh:///root@localhost/system list' > > > > Hi Olivia. > > > > It's in spanish, but pretty easy to follow, maybe it helps you: > > > > > http://wiki.legido.com/doku.php?id=informatica:linux:virtualizacion:kvm#adm > > inistrar_kvm_y_vm_desde_una_gui_en_el_cliente > > > > With this guide you should be able to connect from a client machine (for > > instance your laptop) to the hypervisor (KVM) if both have SSH > connectivity. > > > > Cheers. > > > > Javier > > > > 2013/3/21 Yin Olivia-R63875 <r63875(a)freescale.com>: > > > Hi, > > > > > > I'm trying remote connection with qemu hypervisor on FSL PPC board. > > > > > > The libvirt server is the PPC board. > > > > > > root@ppc:~# ifconfig eth0 10.193.20.109 root@ppc:~# libvirtd -d > > > root@ppc:~# virsh -c qemu:///system define test.xml root@ppc:~# virsh > > > -c qemu:///system start test root@ppc:~# virsh -c qemu:///system list > > > --all > > > Id Name State > > > ---------------------------------------------------- > > > 2 test running > > > > > > Connect from an X86 PC (Ubuntu 10.04) to the PPC board. > > > > > > user@x86:~$ virsh -c qemu+ssh://root@10.193.20.109/system list --all > > > The authenticity of host '10.193.20.109 (10.193.20.109)' can't be > > established. > > > RSA key fingerprint is 2f:56:07:08:da:7d:ac:41:45:57:d2:12:15:19:67:e0. > > > Are you sure you want to continue connecting (yes/no)? yes > > > root(a)10.193.20.109's password: > > > error: failed to connect to the hypervisor > > > error: End of file while reading data: Warning: Permanently added > > '10.193.20.109' (RSA) to the list of known hosts. > > > nc: invalid option -- 'U' > > > BusyBox v1.19.4 (2013-03-08 13:08:18 CST) multi-call binary. > > > > > > Usage: nc [-iN] [-wN] [-l] [-p PORT] [-f FILE|IPADDR PORT] [-e PROG]: > > > Input/output error > > > > > > > > > > > > I tried to verify the remote connection on localhost. But it also > failed > > as below: > > > > > > root@mpc8572ds:~# virsh -c qemu+ssh:///root@localhost/system list > > > --all root@localhost's password: > > > error: failed to connect to the hypervisor > > > error: End of file while reading data: nc: invalid option -- 'U' > > > BusyBox v1.19.4 (2013-03-08 13:08:18 CST) multi-call binary. > > > > > > Usage: nc [-iN] [-wN] [-l] [-p PORT] [-f FILE|IPADDR PORT] [-e PROG]: > > > Input/output error > > > > > > > > > Could anyone give suggestion on this issue? > > > > > > > > > Best Regards, > > > Olivia > > > > > > > > > _______________________________________________ > > > libvirt-users mailing list > > > libvirt-users(a)redhat.com > > > https://www.redhat.com/mailman/listinfo/libvirt-users > > >

1 0

[libvirt-users] Libvirt dead, pid still exists
by SHREE DUTH AWASTHI 21 Mar '13

21 Mar '13

Hi All, Refering to the link below, I got to know that previous versions of Libvirtd had this issue. https://www.redhat.com/archives/libvirt-users/2012-August/msg00104.html We are using libvirt-0.10.0 but we are facing the same issue. Procedure to reproduce : 1. Login to some domain name say CLA-0 using virsh. (i) # virsh list --all Id Name State ---------------------------------------------------- 4 CLA-0 running (ii) # virsh console 4 Connected to domain CLA-0 Escape character is ^] 2. Whenever we exit from the above CLA-0 domain by pressing ctrl-] We are getting the Segmentation fault, however core is not generated but we have debugged it through gdb Program received signal SIGSEGV, Segmentation fault. 0x00007ffff5a96c71 in pthread_mutex_lock () from /lib64/libpthread.so.0 (gdb) bt #0 0x00007ffff5a96c71 in pthread_mutex_lock () from /lib64/libpthread.so.0 #1 0x00007ffff78c8b97 in ?? () from /usr/lib64/libvirt.so.0 #2 0x00007ffff78b6297 in ?? () from /usr/lib64/libvirt.so.0 #3 0x00007ffff7865b65 in virHashRemoveEntry () from /usr/lib64/libvirt.so.0 #4 0x00007ffff78b60d0 in ?? () from /usr/lib64/libvirt.so.0 #5 0x00007ffff78c95f9 in ?? () from /usr/lib64/libvirt.so.0 #6 0x00007ffff78ce711 in virStreamAbort () from /usr/lib64/libvirt.so.0 #7 0x000000000042b546 in ?? () #8 0x000000000042b8f3 in ?? () #9 0x00007ffff78c983a in ?? () from /usr/lib64/libvirt.so.0 #10 0x00007ffff7843cc5 in virEventPollRunOnce () from /usr/lib64/libvirt.so.0 #11 0x00007ffff78428f5 in virEventRunDefaultImpl () from /usr/lib64/libvirt.so.0 #12 0x00007ffff793407d in virNetServerRun () from /usr/lib64/libvirt.so.0 #13 0x000000000040c483 in ?? () #14 0x00007ffff534cbbe in __libc_start_main () from /lib64/libc.so.6 ---Type <return> to continue, or q <return> to quit--- #15 0x000000000040b139 in ?? () #16 0x00007fffffffe7d8 in ?? () #17 0x000000000000001c in ?? () #18 0x0000000000000001 in ?? () #19 0x00007fffffffea73 in ?? () #20 0x0000000000000000 in ?? () After this point libvirtd is dead. # service libvirtd status libvirtd dead but pid file exists >From the mail conversation between you and Andrey we came to know that patch has been provided for this problem. Could you please share us the fix. Thanking you in anticipation. Thanks and Regards, Shree Duth Awasthi.

1 0

[libvirt-users] About live migration with snapshots
by Chiang Hubert 21 Mar '13

21 Mar '13

Hello, I'd like to live migration with snapshots. But it doesn't work. It comes out a message "cannot migrate domain with 1 snapshots" Then I try to trace the code(Libvirt 0.9.8 to 1.0.3), I find out the code in src/qemu/qemu_migration.c @ Line 1395 - 1440 (Libvirt 1.0.3) It will check the VM which has snapshots or not. I just curious about this limitation, why the VM can't live migration with snapshots? What happen if I skip this check? Does it has any suggestion way or virsh command with options to do live migration with snapshots? Thanks a lot. Hubert

3 2

[libvirt-users] SSD Trim needed? Physical Block Device Keep Track Of Guest FS Blocks?
by d hee 21 Mar '13

21 Mar '13

I am trying to figure out if a SSD drive needs to receive trim commands from the KVM guest filesystem(strictly Linux ext4 in this case).... When a file is deleted inside a KVM guest(ext4 and raw image)....If a SSD is the underlying block device that the KVM raw disk image resides on, does the SDD drive need to know which block(s) are marked for reuse within the raw image? Or is it since the Guest raw image is a container that the physical hard drive has no concern about the filesystem blocks within the Guest? Meaning the physical disk would only be concerned with the image itself as a whole...such as if the whole image was deleted? Thanks, -Darin

2 1

[libvirt-users] Storage pool created on the "wrong" vg
by Alaric Haag 19 Mar '13

19 Mar '13

Hello all, (Hopefully, I am posing this question to the correct list!) I seem to have mis-clicked through the creation of an LVM-based storage pool in virt-manager and it is using a volume group containing LVs dedicated to the root/swap filesystems. As these are active LVs, there seems no way to remove this pool. Is that true? My reading thus far suggests that, with LVM storage units, you can't delete them from the pool without REALLY deleting them from the VG. I'm "new enough" to KVM to be missing something obvious, but I'd be pretty shocked if a "structure" like a storage pool can't be deconstructed. Kind regards, Alaric

1 0

[libvirt-users] network bridge hairpin parameter support
by Zang MingJie 19 Mar '13

19 Mar '13

Hi: Currently we encounter a problem where OpenStack may forget or misconfig the bridge hairpin setting, and will cause some unexpected behavior [1]. Although this bug is marked fixed, but if we manually restart the vm by virsh will also trigger the problem. And because the bridge interface is created and managed by libvirt, we are considering move the hairpin configuration to libvirt. Now I start writing a patch to achieve this, but after digging into the source code, I found only kvm/qemu and uml backend may support this parameter. Other backends manage their own bridge, and seems no way to modify it. I expect is there any solution working in progress or should we add the new hairpin parameter in libvirt configuration ? [1] https://bugs.launchpad.net/nova/+bug/933640

1 0