- Devel - libvirt List Archives

[libvirt] Assigning Static IP through libvirt.
by Kumar L Srikanth-B22348 18 Mar '10

18 Mar '10

Hi, I want to assign a static IP address to one of the interfaces created through libvirt. Can anyone please let me know the network XML format? I explored lot of sites on this, but I only found assigning IP address through DHCP rather than Static. Can u please help me. Regards, Srikanth.

1 0

[libvirt] [PATCH 9/13] Add virsh support for new CLI commands
by Stefan Berger 18 Mar '10

18 Mar '10

This patch adds virsh support for the five new CLI commands to manage network filters. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

3 3

[libvirt] [PATCH] docs: <pre> cannot be nested in 
by Matthias Bolte 17 Mar '10

17 Mar '10

xsltproc complained about this. --- docs/hacking.html.in | 7 ++++--- 1 files changed, 4 insertions(+), 3 deletions(-) diff --git a/docs/hacking.html.in b/docs/hacking.html.in index f5ec635..03a1bee 100644 --- a/docs/hacking.html.in +++ b/docs/hacking.html.in @@ -128,20 +128,19 @@ For variadic macros, stick with C99 syntax: + <pre> #define vshPrint(_ctl, ...) fprintf(stdout, __VA_ARGS__) </pre> - Use parenthesis when checking if a macro is defined, and use indentation to track nesting: - + <pre> #if defined(HAVE_POSIX_FALLOCATE) && !defined(HAVE_FALLOCATE) # define fallocate(a,ignored,b,c) posix_fallocate(a,b,c) #endif </pre> - <h2><a href="types">C types</a></h2> @@ -552,9 +551,11 @@ also rebuild locally, run 'make check syntax-check', and make sure you don't raise errors. Try to look for warnings too; for example, configure with + <pre> --enable-compile-warnings=error </pre> + which adds -Werror to compile flags, so no warnings get missed -- 1.6.3.3

2 2

Re: [libvirt] virsh stucked in virDomainCreate
by Matthias Bolte 17 Mar '10

17 Mar '10

[Let's keep it on the list, so other could help too] 2010/3/15 Anthony Lannuzel <lannuzel(a)gmail.com>: > On Sat, Mar 13, 2010 at 6:51 PM, Matthias Bolte > <matthias.bolte(a)googlemail.com> wrote: >> 2010/3/13 Anthony Lannuzel <lannuzel(a)gmail.com>: >>> Hi, >>> >>> I'm having an issue when trying to start a (previously created) vbox >>> item : virsh gets stucked in virDomainCreate and I see no error >>> message (LIBVIRT_DEBUG=1). >>> Is there any mean to get more information on what is happening ? >>> >>> Moreover, now that virsh has stalled, running "connect >>> vbox:///session" on another virsh instance gets an error from "secret >>> driver". Is this normal, as if I'm only supposed to have 1 connection >>> at a time ? >>> >>> Thanks >>> Regards >>> >>> Anthony >>> >> >> I tested starting vbox guest using virsh and it works as expected and >> I can use multiple connections at the same time as expected. >> >> What libvirt and vbox versions do you use? >> >> Can you give some more details about what you did to trigger this problem? >> >> Matthias >> > > Hi, > > I'm using virsh 0.7.6 with VirtualBox 3.1.4 on a 2.6.24 kernel with > openVZ, unionfs and squashfs patches. > > Here is what I do : > > The domain is created with virDomainDefineXML. Here is the XML file content: > > <domain type="vbox"> > <name>c7c0ab9f-9a19-4a45-92a1-799367789509</name> > <uuid>c7c0ab9f-9a19-4a45-92a1-799367789509</uuid> > <vcpu>1</vcpu> > <memory>196608</memory> > <on_poweroff>destroy</on_poweroff> > <on_reboot>destroy</on_reboot> > <on_crash>destroy</on_crash> > <os> > <type>hvm</type> > </os> > <features> > <pae></pae> > <acpi></acpi> > <apic></apic> > </features> > <devices> > <disk device="disk" type="file"> > <source file="/data/c7c0ab9f-9a19-4a45-92a1-799367789509/ubuntuDesktop-8.04.vdi"/> > <target bus="ide" dev="hda"/> > </disk> > <interface type="bridge"> > <source bridge="hnsTap0_tap"/> > <mac address="00:22:94:ae:a1:86"/> > <model type="82543gc"/> > </interface> > <graphics port="12000" type="rdp"/> > </devices> > </domain> > > > > Then I try to start it with virsh: > > root@hynesim-live:/home/hynesim# virsh > 10:09:06.685: debug : virInitialize:336 : register drivers > 10:09:06.685: debug : virRegisterDriver:837 : registering Test as driver 0 > 10:09:06.685: debug : virRegisterNetworkDriver:675 : registering Test > as network driver 0 > 10:09:06.685: debug : virRegisterInterfaceDriver:706 : registering > Test as interface driver 0 > 10:09:06.685: debug : virRegisterStorageDriver:737 : registering Test > as storage driver 0 > 10:09:06.685: debug : virRegisterDeviceMonitor:768 : registering Test > as device driver 0 > 10:09:06.685: debug : virRegisterSecretDriver:799 : registering Test > as secret driver 0 > 10:09:06.685: debug : virRegisterDriver:837 : registering OPENVZ as driver 1 > 10:09:06.709: debug : vboxRegister:87 : VBoxCGlueInit found API > version: 3.1.4 (3001004) > 10:09:06.718: debug : vboxRegister:104 : VirtualBox API version: 3.1 > 10:09:06.718: debug : virRegisterDriver:837 : registering VBOX as driver 2 > 10:09:06.718: debug : virRegisterNetworkDriver:675 : registering VBOX > as network driver 1 > 10:09:06.718: debug : virRegisterStorageDriver:737 : registering VBOX > as storage driver 1 > 10:09:06.718: debug : virRegisterDriver:837 : registering remote as driver 3 > 10:09:06.718: debug : virRegisterNetworkDriver:675 : registering > remote as network driver 2 > 10:09:06.719: debug : virRegisterInterfaceDriver:706 : registering > remote as interface driver 1 > 10:09:06.719: debug : virRegisterStorageDriver:737 : registering > remote as storage driver 2 > 10:09:06.719: debug : virRegisterDeviceMonitor:768 : registering > remote as device driver 1 > 10:09:06.719: debug : virRegisterSecretDriver:799 : registering remote > as secret driver 1 > 10:09:06.719: debug : virConnectOpenAuth:1355 : name=(null), > auth=0x2b78d7561c80, flags=0 > 10:09:06.719: debug : do_open:1112 : no name, allowing driver auto-select > 10:09:06.719: debug : do_open:1120 : trying driver 0 (Test) ... > 10:09:06.719: debug : do_open:1126 : driver 0 Test returned DECLINED > 10:09:06.719: debug : do_open:1120 : trying driver 1 (OPENVZ) ... > 10:09:06.759: debug : virExecWithHook:637 : LC_ALL=C /usr/sbin/vzctl --help > 10:09:06.771: debug : do_open:1126 : driver 1 OPENVZ returned SUCCESS > 10:09:06.771: debug : do_open:1146 : network driver 0 Test returned DECLINED > 10:09:06.771: debug : do_open:1146 : network driver 1 VBOX returned DECLINED > 10:09:06.772: debug : doRemoteOpen:564 : proceeding with name = openvz:///system > 10:09:06.774: debug : remoteIO:8429 : Do proc=66 serial=0 length=28 wait=(nil) > 10:09:06.774: debug : remoteIO:8491 : We have the buck 66 > 0x2b78dbcf2010 0x2b78dbcf2010 > 10:09:06.776: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 64 total (60 more) > 10:09:06.776: debug : remoteIOEventLoop:8355 : Giving up the buck 66 > 0x2b78dbcf2010 (nil) > 10:09:06.776: debug : remoteIO:8522 : All done with our call 66 (nil) > 0x2b78dbcf2010 > 10:09:06.776: debug : remoteIO:8429 : Do proc=1 serial=1 length=56 wait=(nil) > 10:09:06.776: debug : remoteIO:8491 : We have the buck 1 0x63fe50 0x63fe50 > 10:09:06.848: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 56 total (52 more) > 10:09:06.848: debug : remoteIOEventLoop:8355 : Giving up the buck 1 > 0x63fe50 (nil) > 10:09:06.848: debug : remoteIO:8522 : All done with our call 1 (nil) 0x63fe50 > 10:09:06.848: debug : doRemoteOpen:891 : Adding Handler for remote events > 10:09:06.848: debug : doRemoteOpen:898 : virEventAddHandle failed: No > addHandleImpl defined. continuing without events. > 10:09:06.848: debug : do_open:1146 : network driver 2 remote returned SUCCESS > 10:09:06.848: debug : do_open:1165 : interface driver 0 Test returned DECLINED > 10:09:06.849: debug : doRemoteOpen:564 : proceeding with name = openvz:///system > 10:09:06.850: debug : remoteIO:8429 : Do proc=66 serial=0 length=28 wait=(nil) > 10:09:06.850: debug : remoteIO:8491 : We have the buck 66 0x67ff00 0x67ff00 > 10:09:06.852: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 64 total (60 more) > 10:09:06.852: debug : remoteIOEventLoop:8355 : Giving up the buck 66 > 0x67ff00 (nil) > 10:09:06.852: debug : remoteIO:8522 : All done with our call 66 (nil) 0x67ff00 > 10:09:06.853: debug : remoteIO:8429 : Do proc=1 serial=1 length=56 wait=(nil) > 10:09:06.853: debug : remoteIO:8491 : We have the buck 1 0x67ff00 0x67ff00 > 10:09:06.901: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 56 total (52 more) > 10:09:06.901: debug : remoteIOEventLoop:8355 : Giving up the buck 1 > 0x67ff00 (nil) > 10:09:06.901: debug : remoteIO:8522 : All done with our call 1 (nil) 0x67ff00 > 10:09:06.901: debug : doRemoteOpen:891 : Adding Handler for remote events > 10:09:06.901: debug : doRemoteOpen:898 : virEventAddHandle failed: No > addHandleImpl defined. continuing without events. > 10:09:06.901: debug : do_open:1165 : interface driver 1 remote returned SUCCESS > 10:09:06.901: debug : do_open:1185 : storage driver 0 Test returned DECLINED > 10:09:06.901: debug : do_open:1185 : storage driver 1 VBOX returned DECLINED > 10:09:06.901: debug : do_open:1185 : storage driver 2 remote returned SUCCESS > 10:09:06.901: debug : do_open:1205 : node driver 0 Test returned DECLINED > 10:09:06.901: debug : do_open:1205 : node driver 1 remote returned SUCCESS > 10:09:06.901: debug : do_open:1232 : secret driver 0 Test returned DECLINED > 10:09:06.901: debug : do_open:1232 : secret driver 1 remote returned SUCCESS > Bienvenue dans virsh, le terminal de virtualisation interactif. > > Taper : « help » pour l'aide ou « help » avec la commande > « quit » pour quitter > > virsh # connect vbox:///session > 10:09:13.993: debug : virConnectClose:1381 : conn=0x63b060 > 10:09:13.994: debug : virUnrefConnect:259 : unref connection 0x63b060 1 > 10:09:13.994: debug : remoteIO:8429 : Do proc=2 serial=2 length=28 wait=(nil) > 10:09:13.994: debug : remoteIO:8491 : We have the buck 2 0x680170 0x680170 > 10:09:13.995: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 56 total (52 more) > 10:09:13.995: debug : remoteIOEventLoop:8355 : Giving up the buck 2 > 0x680170 (nil) > 10:09:13.995: debug : remoteIO:8522 : All done with our call 2 (nil) 0x680170 > 10:09:13.996: debug : remoteIO:8429 : Do proc=2 serial=2 length=28 wait=(nil) > 10:09:13.996: debug : remoteIO:8491 : We have the buck 2 0x680170 0x680170 > 10:09:13.997: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 56 total (52 more) > 10:09:13.997: debug : remoteIOEventLoop:8355 : Giving up the buck 2 > 0x680170 (nil) > 10:09:13.997: debug : remoteIO:8522 : All done with our call 2 (nil) 0x680170 > 10:09:13.997: debug : virDomainObjUnref:680 : obj=0x63ebf0 refs=0 > 10:09:13.997: debug : virDomainObjFree:658 : obj=0x63ebf0 > 10:09:13.997: debug : virReleaseConnect:216 : release connection 0x63b060 > 10:09:13.997: debug : virConnectOpenAuth:1355 : name=vbox:///session, > auth=0x2b78d7561c80, flags=0 > 10:09:13.997: debug : do_open:1110 : name "vbox:///session" to URI components: > scheme vbox > opaque (null) > authority (null) > server (null) > user (null) > port 0 > path /session > > 10:09:13.997: debug : do_open:1120 : trying driver 0 (Test) ... > 10:09:13.997: debug : do_open:1126 : driver 0 Test returned DECLINED > 10:09:13.997: debug : do_open:1120 : trying driver 1 (OPENVZ) ... > 10:09:13.997: debug : do_open:1126 : driver 1 OPENVZ returned DECLINED > 10:09:13.997: debug : do_open:1120 : trying driver 2 (VBOX) ... > 10:09:14.021: debug : vboxOpen:836 : in vboxOpen > 10:09:14.021: debug : do_open:1126 : driver 2 VBOX returned SUCCESS > 10:09:14.021: debug : do_open:1146 : network driver 0 Test returned DECLINED > 10:09:14.021: debug : vboxNetworkOpen:5434 : network initialized > 10:09:14.021: debug : do_open:1146 : network driver 1 VBOX returned SUCCESS > 10:09:14.022: debug : do_open:1165 : interface driver 0 Test returned DECLINED > 10:09:14.022: debug : doRemoteOpen:564 : proceeding with name = vbox:///session > 10:09:14.024: debug : remoteIO:8429 : Do proc=66 serial=0 length=28 wait=(nil) > 10:09:14.024: debug : remoteIO:8491 : We have the buck 66 0x6eaa90 0x6eaa90 > 10:09:14.025: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 64 total (60 more) > 10:09:14.025: debug : remoteIOEventLoop:8355 : Giving up the buck 66 > 0x6eaa90 (nil) > 10:09:14.025: debug : remoteIO:8522 : All done with our call 66 (nil) 0x6eaa90 > 10:09:14.026: debug : remoteIO:8429 : Do proc=1 serial=1 length=56 wait=(nil) > 10:09:14.026: debug : remoteIO:8491 : We have the buck 1 0x6eaa90 0x6eaa90 > 10:09:14.027: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 56 total (52 more) > 10:09:14.027: debug : remoteIOEventLoop:8355 : Giving up the buck 1 > 0x6eaa90 (nil) > 10:09:14.027: debug : remoteIO:8522 : All done with our call 1 (nil) 0x6eaa90 > 10:09:14.027: debug : doRemoteOpen:891 : Adding Handler for remote events > 10:09:14.027: debug : doRemoteOpen:898 : virEventAddHandle failed: No > addHandleImpl defined. continuing without events. > 10:09:14.027: debug : do_open:1165 : interface driver 1 remote returned SUCCESS > 10:09:14.027: debug : do_open:1185 : storage driver 0 Test returned DECLINED > 10:09:14.027: debug : vboxStorageOpen:6186 : vbox storage initialized > 10:09:14.027: debug : do_open:1185 : storage driver 1 VBOX returned SUCCESS > 10:09:14.028: debug : do_open:1205 : node driver 0 Test returned DECLINED > 10:09:14.028: debug : doRemoteOpen:564 : proceeding with name = vbox:///session > 10:09:14.030: debug : remoteIO:8429 : Do proc=66 serial=0 length=28 wait=(nil) > 10:09:14.037: debug : remoteIO:8491 : We have the buck 66 0x72ab40 0x72ab40 > 10:09:14.038: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 64 total (60 more) > 10:09:14.038: debug : remoteIOEventLoop:8355 : Giving up the buck 66 > 0x72ab40 (nil) > 10:09:14.038: debug : remoteIO:8522 : All done with our call 66 (nil) 0x72ab40 > 10:09:14.038: debug : remoteIO:8429 : Do proc=1 serial=1 length=56 wait=(nil) > 10:09:14.039: debug : remoteIO:8491 : We have the buck 1 0x72ab40 0x72ab40 > 10:09:14.040: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 56 total (52 more) > 10:09:14.040: debug : remoteIOEventLoop:8355 : Giving up the buck 1 > 0x72ab40 (nil) > 10:09:14.040: debug : remoteIO:8522 : All done with our call 1 (nil) 0x72ab40 > 10:09:14.040: debug : doRemoteOpen:891 : Adding Handler for remote events > 10:09:14.040: debug : doRemoteOpen:898 : virEventAddHandle failed: No > addHandleImpl defined. continuing without events. > 10:09:14.040: debug : do_open:1205 : node driver 1 remote returned SUCCESS > 10:09:14.040: debug : do_open:1232 : secret driver 0 Test returned DECLINED > 10:09:14.040: debug : doRemoteOpen:564 : proceeding with name = vbox:///session > 10:09:14.041: debug : remoteIO:8429 : Do proc=66 serial=0 length=28 wait=(nil) > 10:09:14.041: debug : remoteIO:8491 : We have the buck 66 0x76abf0 0x76abf0 > 10:09:14.042: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 64 total (60 more) > 10:09:14.042: debug : remoteIOEventLoop:8355 : Giving up the buck 66 > 0x76abf0 (nil) > 10:09:14.042: debug : remoteIO:8522 : All done with our call 66 (nil) 0x76abf0 > 10:09:14.043: debug : remoteIO:8429 : Do proc=1 serial=1 length=56 wait=(nil) > 10:09:14.043: debug : remoteIO:8491 : We have the buck 1 0x76abf0 0x76abf0 > 10:09:14.044: debug : remoteIODecodeMessageLength:7913 : Got length, > now need 56 total (52 more) > 10:09:14.044: debug : remoteIOEventLoop:8355 : Giving up the buck 1 > 0x76abf0 (nil) > 10:09:14.044: debug : remoteIO:8522 : All done with our call 1 (nil) 0x76abf0 > 10:09:14.044: debug : doRemoteOpen:891 : Adding Handler for remote events > 10:09:14.044: debug : doRemoteOpen:898 : virEventAddHandle failed: No > addHandleImpl defined. continuing without events. > 10:09:14.044: debug : do_open:1232 : secret driver 1 remote returned SUCCESS > > virsh # list --all > 10:09:19.563: debug : virConnectNumOfDomains:1749 : conn=0x63ebf0 > 10:09:19.568: debug : virConnectNumOfDefinedDomains:4609 : conn=0x63ebf0 > 10:09:19.568: debug : virConnectListDefinedDomains:4648 : > conn=0x63ebf0, names=0x692430, maxnames=1 > ID Nom État > ---------------------------------- > 10:09:19.570: debug : virDomainLookupByName:2010 : conn=0x63ebf0, > name=c7c0ab9f-9a19-4a45-92a1-799367789509 > 10:09:19.571: debug : virGetDomain:345 : New hash entry 0x6a6000 > 10:09:19.571: debug : virDomainGetInfo:2867 : domain=0x6a6000, > info=0x7fffd3a519e0 > - c7c0ab9f-9a19-4a45-92a1-799367789509 closed > 10:09:19.573: debug : virDomainFree:2098 : domain=0x6a6000 > 10:09:19.573: debug : virUnrefDomain:422 : unref domain 0x6a6000 > c7c0ab9f-9a19-4a45-92a1-799367789509 1 > 10:09:19.573: debug : virReleaseDomain:376 : release domain 0x6a6000 > c7c0ab9f-9a19-4a45-92a1-799367789509 > 10:09:19.573: debug : virReleaseDomain:392 : unref connection 0x63ebf0 2 > > virsh # start c7c0ab9f-9a19-4a45-92a1-799367789509 > 10:09:26.108: debug : virDomainLookupByName:2010 : conn=0x63ebf0, > name=c7c0ab9f-9a19-4a45-92a1-799367789509 > 10:09:26.112: debug : virGetDomain:345 : New hash entry 0x68bd70 > 10:09:26.112: debug : virDomainGetID:2658 : domain=0x68bd70 > 10:09:26.112: debug : virDomainCreate:4690 : domain=0x68bd70 > 10:09:26.112: debug : virDomainCreate:4693 : 1 0x68bd70 > 10:09:26.112: debug : virDomainCreate:4700 : 2 0x68bd70 > 10:09:26.112: debug : virDomainCreate:4707 : 3 0x68bd70 > 10:09:26.112: debug : virDomainCreate:4710 : 3.1 0x68bd70 > 10:09:26.223: debug : vboxDomainCreate:3321 : before close 0x68bd70 > > > I have added some debug lines, that lead to: > > in libvirt.c: > function int virDomainCreate(virDomainPtr domain) > ret = conn->driver->domainCreate (domain); // does not return > > in vbox/vbox_tmpl.c: > function static int vboxDomainCreate(virDomainPtr dom) > data->vboxSession->vtbl->Close(data->vboxSession); > // does not return > > Thanks > At that point the domain should already be running. I wonder why closing the session handle would block. You could try variations of the domain XML config, assuming that a specify part of the config triggers this problem. For example use your posted domain XML config, but leave out the graphics part or the interface part or the disk part to see if that makes a difference. Matthias

1 0

[libvirt] Userdata for libvirt?
by Yushu Yao 17 Mar '10

17 Mar '10

Hi Experts, Forgive me if not clear enough, will try to be describe what I need. What I need: Start a VM on a remote libvirtd server based on an existing image (Might or might not exist on server), plus some userdata. It is pretty much the same like EC2 or Eucalyptus, but we need to run on our private cluster with only libvirtd. (Eucalyptus doesn't support using physical disk in VMs, that is what we need). I went through the Remote Libvirt control section of the Documentation, but couldn't find a way to pass the userdata to the VM started. Does libvirtd already include such feature of do I need to make my own. Thanks -Yushu

2 2

[libvirt] Just pushed a few small fixes
by Cole Robinson 17 Mar '10

17 Mar '10

Hi all, I've just pushed three small fixes: $ git log -3 --pretty=short commit 2ef091efcc5cd02bbd496972b141cf253f713fdd Author: Philip Hahn <hahn(a)univention.de> python: Fix networkLookupByUUID commit 0ef58c315506d094a0a4d64f022119a04a1916c0 Author: Cole Robinson <crobinso(a)redhat.com> .gitignore: Ignore generated daemon/libvirtd.logrotate commit 89d8cdfc7e0ba7b2d9cf16a34b8e4a035c851d28 Author: Cole Robinson <crobinso(a)redhat.com> Fix make dist with XenAPI changes The last chronological change was from: https://bugzilla.redhat.com/show_bug.cgi?id=574366 Apparently networkLookupByUUID has always been busted in the python bindings, but since we don't use it in virtinst/virt-manager, it wasn't noticed. - Cole

1 0

[libvirt] disk XML for error policy
by Dave Allan 17 Mar '10

17 Mar '10

Hi Dan, What do you think of this for the disk XML specifying what to do on error: <disk type='file' device='disk'> <source file='/storage/guest_disks/libgdu'/> <target dev='vda' bus='virtio'/> <on_err rerror='stop' werror='stop'/> </disk> Dave

2 2

[libvirt] [PATCH] do not require two ./autogen.sh runs to permit "make"
by Jim Meyering 17 Mar '10

17 Mar '10

I tracked down the source of the two-autogen.sh-runs-required bug. Here's the fix: >From 0583de2de2038d4f8875268391fd4994329d39bf Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 16 Mar 2010 21:08:31 +0100 Subject: [PATCH] do not require two ./autogen.sh runs to permit "make" * autogen.sh (bootstrap_hash): New function. Running bootstrap may update the gnulib SHA1, yet we were computing t=$(git submodule status ...) *prior* to running bootstrap, and then recording that sometimes-stale value in the stamp file upon a successful bootstrap run. That would require two (lengthy!) bootstrap runs to update the stamp file. --- autogen.sh | 21 ++++++++++++++------- 1 files changed, 14 insertions(+), 7 deletions(-) diff --git a/autogen.sh b/autogen.sh index ff94678..b93cdba 100755 --- a/autogen.sh +++ b/autogen.sh @@ -62,20 +62,27 @@ else fi fi +# Compute the hash we'll use to determine whether rerunning bootstrap +# is required. The first is just the SHA1 that selects a gnulib snapshot. +# The second ensures that whenever we change the set of gnulib modules used +# by this package, we rerun bootstrap to pull in the matching set of files. +bootstrap_hash() +{ + git submodule status | sed 's/^[ +-]//;s/ .*//' + git hash-object bootstrap.conf +} + # Ensure that whenever we pull in a gnulib update or otherwise change to a # different version (i.e., when switching branches), we also rerun ./bootstrap. curr_status=.git-module-status -t=$(git submodule status|sed 's/^[ +-]//;s/ .*//'; \ - git hash-object bootstrap.conf) +t=$(bootstrap_hash) if test "$t" = "$(cat $curr_status 2>/dev/null)"; then : # good, it's up to date, all we need is autoreconf autoreconf -if else - echo running bootstrap... - ./bootstrap && echo "$t" > $curr_status || { - echo "Failed to bootstrap gnulib, please investigate." - exit 1; - } + echo running bootstrap... + ./bootstrap && bootstrap_hash > $curr_status \ + || { echo "Failed to bootstrap gnulib, please investigate."; exit 1; } fi cd "$THEDIR" -- 1.7.0.1

3 4

[libvirt] [PATCH] Allow suspend during live migration
by Jiri Denemark 17 Mar '10

17 Mar '10

Currently no command can be sent to a qemu process while another job is active. This patch adds support for signaling long-running jobs (such as migration) so that other threads may request predefined operations to be done during such jobs. Two signals are defined so far: - QEMU_JOB_SIGNAL_CANCEL - QEMU_JOB_SIGNAL_SUSPEND The first one is used by qemuDomainAbortJob. The second one is used by qemudDomainSuspend for suspending a domain during migration, which allows for changing live migration into offline migration. However, there is a small issue in the way qemudDomainSuspend is currently implemented for migrating domains. The API calls returns immediately after signaling migration job which means it is asynchronous in this specific case. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/qemu/qemu_driver.c | 149 ++++++++++++++++++++++++++++++++--------------- 1 files changed, 101 insertions(+), 48 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index f8ab545..5b2c26a 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -87,14 +87,26 @@ #define VIR_FROM_THIS VIR_FROM_QEMU +/* Only 1 job is allowed at any time + * A job includes *all* monitor commands, even those just querying + * information, not merely actions */ +enum qemuDomainJob { + QEMU_JOB_NONE = 0, /* Always set to 0 for easy if (jobActive) conditions */ + QEMU_JOB_UNSPECIFIED, + QEMU_JOB_MIGRATION, +}; + +enum qemuDomainJobSignals { + QEMU_JOB_SIGNAL_CANCEL = 1 << 0, /* Request job cancellation */ + QEMU_JOB_SIGNAL_SUSPEND = 1 << 1, /* Request VM suspend to finish live migration offline */ +}; + typedef struct _qemuDomainObjPrivate qemuDomainObjPrivate; typedef qemuDomainObjPrivate *qemuDomainObjPrivatePtr; struct _qemuDomainObjPrivate { virCond jobCond; /* Use in conjunction with main virDomainObjPtr lock */ - unsigned int jobActive : 1; /* Non-zero if a job is active. Only 1 job is allowed at any time - * A job includes *all* monitor commands, even those just querying - * information, not merely actions */ - unsigned int jobCancel : 1; /* Non-zero if a cancel request from client has arrived */ + enum qemuDomainJob jobActive; /* Currently running job */ + unsigned int jobSignals; /* Signals for running job */ virDomainJobInfo jobInfo; unsigned long long jobStart; @@ -338,8 +350,8 @@ static int qemuDomainObjBeginJob(virDomainObjPtr obj) return -1; } } - priv->jobActive = 1; - priv->jobCancel = 0; + priv->jobActive = QEMU_JOB_UNSPECIFIED; + priv->jobSignals = 0; priv->jobStart = (now.tv_sec * 1000ull) + (now.tv_usec / 1000); memset(&priv->jobInfo, 0, sizeof(priv->jobInfo)); @@ -385,8 +397,8 @@ static int qemuDomainObjBeginJobWithDriver(struct qemud_driver *driver, return -1; } } - priv->jobActive = 1; - priv->jobCancel = 0; + priv->jobActive = QEMU_JOB_UNSPECIFIED; + priv->jobSignals = 0; priv->jobStart = (now.tv_sec * 1000ull) + (now.tv_usec / 1000); memset(&priv->jobInfo, 0, sizeof(priv->jobInfo)); @@ -410,8 +422,8 @@ static int ATTRIBUTE_RETURN_CHECK qemuDomainObjEndJob(virDomainObjPtr obj) { qemuDomainObjPrivatePtr priv = obj->privateData; - priv->jobActive = 0; - priv->jobCancel = 0; + priv->jobActive = QEMU_JOB_NONE; + priv->jobSignals = 0; priv->jobStart = 0; memset(&priv->jobInfo, 0, sizeof(priv->jobInfo)); virCondSignal(&priv->jobCond); @@ -3560,6 +3572,7 @@ static int qemudDomainSuspend(virDomainPtr dom) { virDomainObjPtr vm; int ret = -1; virDomainEventPtr event = NULL; + qemuDomainObjPrivatePtr priv; qemuDriverLock(driver); vm = virDomainFindByUUID(&driver->domains, dom->uuid); @@ -3571,30 +3584,48 @@ static int qemudDomainSuspend(virDomainPtr dom) { _("no domain with matching uuid '%s'"), uuidstr); goto cleanup; } - if (qemuDomainObjBeginJobWithDriver(driver, vm) < 0) - goto cleanup; - if (!virDomainObjIsActive(vm)) { qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", _("domain is not running")); - goto endjob; + goto cleanup; } - if (vm->state != VIR_DOMAIN_PAUSED) { - qemuDomainObjPrivatePtr priv = vm->privateData; - qemuDomainObjEnterMonitorWithDriver(driver, vm); - if (qemuMonitorStopCPUs(priv->mon) < 0) { - qemuDomainObjExitMonitorWithDriver(driver, vm); + + priv = vm->privateData; + + if (priv->jobActive == QEMU_JOB_MIGRATION) { + if (vm->state != VIR_DOMAIN_PAUSED) { + VIR_DEBUG("Requesting domain pause on %s", + vm->def->name); + priv->jobSignals |= QEMU_JOB_SIGNAL_SUSPEND; + } + ret = 0; + goto cleanup; + } else { + if (qemuDomainObjBeginJobWithDriver(driver, vm) < 0) + goto cleanup; + + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); goto endjob; } - qemuDomainObjExitMonitorWithDriver(driver, vm); - vm->state = VIR_DOMAIN_PAUSED; - event = virDomainEventNewFromObj(vm, - VIR_DOMAIN_EVENT_SUSPENDED, - VIR_DOMAIN_EVENT_SUSPENDED_PAUSED); + if (vm->state != VIR_DOMAIN_PAUSED) { + int rc; + + qemuDomainObjEnterMonitorWithDriver(driver, vm); + rc = qemuMonitorStopCPUs(priv->mon); + qemuDomainObjExitMonitorWithDriver(driver, vm); + if (rc < 0) + goto endjob; + vm->state = VIR_DOMAIN_PAUSED; + event = virDomainEventNewFromObj(vm, + VIR_DOMAIN_EVENT_SUSPENDED, + VIR_DOMAIN_EVENT_SUSPENDED_PAUSED); + } + if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) + goto endjob; + ret = 0; } - if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) - goto endjob; - ret = 0; endjob: if (qemuDomainObjEndJob(vm) == 0) @@ -3946,6 +3977,35 @@ cleanup: } +/** qemuDomainMigrateOffline: + * Pause domain for non-live migration. + */ +static int +qemuDomainMigrateOffline(struct qemud_driver *driver, + virDomainObjPtr vm) +{ + qemuDomainObjPrivatePtr priv = vm->privateData; + int ret; + + qemuDomainObjEnterMonitorWithDriver(driver, vm); + ret = qemuMonitorStopCPUs(priv->mon); + qemuDomainObjExitMonitorWithDriver(driver, vm); + + if (ret == 0) { + virDomainEventPtr event; + + vm->state = VIR_DOMAIN_PAUSED; + event = virDomainEventNewFromObj(vm, + VIR_DOMAIN_EVENT_SUSPENDED, + VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED); + if (event) + qemuDomainEventQueue(driver, event); + } + + return ret; +} + + static int qemuDomainWaitForMigrationComplete(struct qemud_driver *driver, virDomainObjPtr vm) { @@ -3964,8 +4024,8 @@ qemuDomainWaitForMigrationComplete(struct qemud_driver *driver, virDomainObjPtr struct timeval now; int rc; - if (priv->jobCancel) { - priv->jobCancel = 0; + if (priv->jobSignals & QEMU_JOB_SIGNAL_CANCEL) { + priv->jobSignals ^= QEMU_JOB_SIGNAL_CANCEL; VIR_DEBUG0("Cancelling migration at client request"); qemuDomainObjEnterMonitorWithDriver(driver, vm); rc = qemuMonitorMigrateCancel(priv->mon); @@ -3973,6 +4033,11 @@ qemuDomainWaitForMigrationComplete(struct qemud_driver *driver, virDomainObjPtr if (rc < 0) { VIR_WARN0("Unable to cancel migration"); } + } else if (priv->jobSignals & QEMU_JOB_SIGNAL_SUSPEND) { + priv->jobSignals ^= QEMU_JOB_SIGNAL_SUSPEND; + VIR_DEBUG0("Pausing domain for non-live migration"); + if (qemuDomainMigrateOffline(driver, vm) < 0) + VIR_WARN0("Unable to pause domain"); } qemuDomainObjEnterMonitorWithDriver(driver, vm); @@ -8979,7 +9044,7 @@ qemudDomainMigratePerform (virDomainPtr dom, virDomainObjPtr vm; virDomainEventPtr event = NULL; int ret = -1; - int paused = 0; + int resume = 0; qemuDomainObjPrivatePtr priv; qemuDriverLock(driver); @@ -8995,6 +9060,7 @@ qemudDomainMigratePerform (virDomainPtr dom, if (qemuDomainObjBeginJobWithDriver(driver, vm) < 0) goto cleanup; + priv->jobActive = QEMU_JOB_MIGRATION; if (!virDomainObjIsActive(vm)) { qemuReportError(VIR_ERR_OPERATION_INVALID, @@ -9005,23 +9071,10 @@ qemudDomainMigratePerform (virDomainPtr dom, memset(&priv->jobInfo, 0, sizeof(priv->jobInfo)); priv->jobInfo.type = VIR_DOMAIN_JOB_UNBOUNDED; + resume = vm->state == VIR_DOMAIN_RUNNING; if (!(flags & VIR_MIGRATE_LIVE) && vm->state == VIR_DOMAIN_RUNNING) { - /* Pause domain for non-live migration */ - qemuDomainObjEnterMonitorWithDriver(driver, vm); - if (qemuMonitorStopCPUs(priv->mon) < 0) { - qemuDomainObjExitMonitorWithDriver(driver, vm); + if (qemuDomainMigrateOffline(driver, vm) < 0) goto endjob; - } - qemuDomainObjExitMonitorWithDriver(driver, vm); - paused = 1; - - vm->state = VIR_DOMAIN_PAUSED; - event = virDomainEventNewFromObj(vm, - VIR_DOMAIN_EVENT_SUSPENDED, - VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED); - if (event) - qemuDomainEventQueue(driver, event); - event = NULL; } if ((flags & (VIR_MIGRATE_TUNNELLED | VIR_MIGRATE_PEER2PEER))) { @@ -9035,7 +9088,7 @@ qemudDomainMigratePerform (virDomainPtr dom, /* Clean up the source domain. */ qemudShutdownVMDaemon(driver, vm); - paused = 0; + resume = 0; event = virDomainEventNewFromObj(vm, VIR_DOMAIN_EVENT_STOPPED, @@ -9049,7 +9102,7 @@ qemudDomainMigratePerform (virDomainPtr dom, ret = 0; endjob: - if (paused) { + if (resume && vm->state == VIR_DOMAIN_PAUSED) { /* we got here through some sort of failure; start the domain again */ qemuDomainObjEnterMonitorWithDriver(driver, vm); if (qemuMonitorStartCPUs(priv->mon, dom->conn) < 0) { @@ -9425,7 +9478,7 @@ static int qemuDomainAbortJob(virDomainPtr dom) { if (virDomainObjIsActive(vm)) { if (priv->jobActive) { VIR_DEBUG("Requesting cancellation of job on vm %s", vm->def->name); - priv->jobCancel = 1; + priv->jobSignals |= QEMU_JOB_SIGNAL_CANCEL; } else { qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", _("no job is active on the domain")); -- 1.7.0.2

2 2

[libvirt] [PATCH 1/13] Adding recursive locks
by Stefan Berger 17 Mar '10

17 Mar '10

This patch adds recursive locks necessary due to the processing of network filter XML that can reference other network filters, including references that cause looks. Loops in the XML are prevented but their detection requires recursive locks.

2 2

[libvirt] [RFC][PATCH v2 0/2] Dynamic backend setup for macvtap interfaces
by Ed Swierk 17 Mar '10

17 Mar '10

I posted this RFC/patch set a few weeks ago but didn't receive any response. My implementation works, but I'd like to hear from anyone more familiar with libvirt concurrency than I (i.e. nearly everyone) about how it might be improved. In v2 I rebased against libvirt-0.7.7 and made a bunch of minor changes. - Using a bridge to connect a qemu NIC to a host interface offers a fair amount of flexibility to reconfigure the host without restarting the VM. For example, if the bridge connects host interface eth0 to the qemu tap0 interface, eth0 can be hot-removed and hot-plugged without affecting the VM. Similarly, if the bridge connects host VLAN interface vlan0 to the qemu tap0 interface, the admin can easily replace vlan0 with vlan1 without the VM noticing. Using the macvtap driver instead of a kernel bridge, the host interface is much more tightly tied to the VM. Qemu communicates with the macvtap interface through a file descriptor, and the macvtap interface is bound permanently to a specific host interface when it is first created. What's more, if the underlying host interface disappears, the macvtap interface vanishes along with it, leaving the VM holding a file descriptor for a deleted file. To avoid race conditions during system startup, I would like libvirt to allow starting up the VM with a NIC even if the underlying host interface doesn't yet exist, deferring creation of the macvtap interface (analogous to starting up the VM with a tap interface bound to an orphan bridge). To support adding and removing a host interface without restarting the VM, I would like libvirt to react to the (re)appearance of the underlying host interface, creating a new macvtap interface and passing the new fd to qemu to reconnect to the NIC. (It would also be nice if libvirt allowed the user to change which underlying host interface the qemu NIC is connected to. I'm ignoring this issue for now, except to note that implementing the above features should make this easier.) The libvirt API already supports domainAttachDevice and domainDetachDevice to add or remove an interface while the VM is running. In the qemu implementation, these commands add or remove the VM NIC device as well as reconfiguring the host side. This works only if the OS and application running in the VM can handle PCI hotplug and dynamically reconfigure its network. I would like to isolate the VM from changes to the host network setup, whether you use macvtap or a bridge. The changes I think are needed to implement this include: 1. Refactor qemudDomainAttachNetDevice/qemudDomainDetachNetDevice, which currently handle both backend (host) setup and adding/removing the VM NIC device; move the backend setup code into separate functions that can called separately without affecting VM devices. 2. Implement a thread or task that watches for changes to the underlying host interface for each configured macvtap interface, and reacts by invoking the appropriate backend setup code. 3. Change qemudBuildCommandLine to defer backend setup if qemu supports the necessary features for doing it later (e.g. the host_net_add monitor command). 4. Implement appropriate error handling and reporting, and any necessary changes to the configuration schema. The following patches are a partial implementation of the above as a proof of concept. Patch 1 implements change (1) above, moving the backend setup code to new functions qemudDomainConnectNetBackend/qemudDomainDisconnectNetBackend, and calling these functions from the existing qemudDomainAttachNetDevice/qemudDomainDetachNetDevice. I think this change is useful on its own: it breaks up two monster functions into more manageable pieces, and eliminates some code duplication (e.g. the try_remove clause at the end of qemudDomainAttachNetDevice). Patch 2 is a godawful hack roughly implementing changes (2) and (3) above (did I mention that this is a proof of concept?). It spawns a thread that simply tries reconnecting the backend of each macvtap interface once a second. As long as the interface is already up, the reconnection fails. If the macvtap interface goes away because the underlying host interface disappears, the reconnection fails until the host interface reappears. I ran into two major issues while implementing (2) and (3): - Can we use the existing virEvent functions to invoke the reconnection process, triggered either by a timer or by an event from the host? It seems like this ought to work, but it appears that communication between libvirt and the qemu monitor relies on an event, and since all events run in the same thread, there's no way for an event to call the monitor. - Should the reconnection process use udev or hal to get notifications, or leverage the node device code which itself uses udev or hal? Currently there doesn't appear to be a way to get notifications of changes to node devices; if there were, we'd still need to address the threading issue. If we use node devices, what changes to the configuration schema would be needed to associate a macvtap interface with the underlying node device? I'd appreciate input on item (4) as well (e.g. does it always make sense to ignore the missing host interface on the assumption that it could show up later?). --Ed

3 3

[libvirt] [PATCH] fix two "make syntax check" failures
by Jim Meyering 16 Mar '10

16 Mar '10

Laine mentioned that there were two syntax-check failures. I've just pushed this fix: >From a31bc6750347a79dbd43b0aacecf86d204e6e160 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 16 Mar 2010 19:32:05 +0100 Subject: [PATCH] fix two "make syntax check" failures * src/xenapi/xenapi_driver.c (xenapiOpen): Remove useless-if-before-free. * po/POTFILES.in: Add src/xenapi/xenapi_utils.c. --- po/POTFILES.in | 1 + src/xenapi/xenapi_driver.c | 2 +- 2 files changed, 2 insertions(+), 1 deletions(-) diff --git a/po/POTFILES.in b/po/POTFILES.in index ac169be..1845572 100644 --- a/po/POTFILES.in +++ b/po/POTFILES.in @@ -78,5 +78,6 @@ src/xen/xen_hypervisor.c src/xen/xen_inotify.c src/xen/xm_internal.c src/xen/xs_internal.c +src/xenapi/xenapi_utils.c tools/console.c tools/virsh.c diff --git a/src/xenapi/xenapi_driver.c b/src/xenapi/xenapi_driver.c index 153582d..ad77068 100644 --- a/src/xenapi/xenapi_driver.c +++ b/src/xenapi/xenapi_driver.c @@ -102,7 +102,7 @@ xenapiOpen (virConnectPtr conn, virConnectAuthPtr auth, int flags ATTRIBUTE_UNUS } if (!passwd || !conn->uri->user) { xenapiSessionErrorHandler(conn, VIR_ERR_AUTH_FAILED, "Username/Password not valid"); - if (passwd) VIR_FREE(passwd); + VIR_FREE(passwd); return VIR_DRV_OPEN_ERROR; } if (VIR_ALLOC(privP) < 0) { -- 1.7.0.2.445.g0ae494

1 0

[libvirt] [PATCH] Use WARN_CFLAGS when compiling virsh.c
by Jiri Denemark 16 Mar '10

16 Mar '10

Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- tools/Makefile.am | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/tools/Makefile.am b/tools/Makefile.am index 941e93e..46107f6 100644 --- a/tools/Makefile.am +++ b/tools/Makefile.am @@ -48,6 +48,7 @@ virsh_CFLAGS = \ -I$(top_srcdir)/src/util \ -DGETTEXT_PACKAGE=\"$(PACKAGE)\" \ -DLOCALEBASEDIR=\""$(datadir)/locale"\" \ + $(WARN_CFLAGS) \ $(COVERAGE_CFLAGS) \ $(LIBXML_CFLAGS) \ $(READLINE_CFLAGS) -- 1.7.0.2

2 2

[libvirt] Fix slow storage file allocation with O_DSYNC
by Jiri Denemark 16 Mar '10

16 Mar '10

Hi. By opening storage file with O_DSYNC before allocating disk blocks for it we made the whole operation terribly slow when the allocation is done using posix_fallocate() on a filesystem which does not support fallocate(), i.e. anything but ext4. And by terribly slow I mean 45 minutes (RHEL5) to 10+ hours (Fedora Rawhide) for a 12GB storage. To fix this issue, we have two options. Either avoid using fallocate() emulation inside posix_fallocate() or stop using O_DSYNC. I prepared a patch for each of these options and they will follow as replies to this introduction. Some numbers for the same 12GB file: - libvirt's internal fallocate emulation using - mmap ~3.5 minutes (will be used whenever mmap() is present) - write ~5 minutes - fsync() instead of O_DSYNC - ~3.5 minutes So basically the two options are equivalent wrt to time consumption. The second one seems to be less hacky at the expense of non-lienar behavior which makes progress reporting less useful. Jirka

3 11

[libvirt] 'build' on FS pool now unconditionally formats?
by Cole Robinson 16 Mar '10

16 Mar '10

Hi guys, Looking at the new FS pool build options and talking with Dave, I see that calling PoolBuild on an FS pool now unconditionally calls mkfs. This is really bad when mixed with virt-manager: previously, we assumed the FS build command was always non destructive (at most it created a directory), so we called it every time, and didn't even allow users to opt out, since there wasn't a use case that called for it. This new formatting behavior really needs to be opt in, otherwise all virt-manager versions creating an FS pool can destroy data. Just FYI, for disk pools (and certain LVM configurations) where this operation has always been destructive, we default to build=off, and loudly warn the user if they choose otherwise. We can do that with this new option as well, but the previous behavior really needs to be reinstated IMO (and before the new release). I fully accept that this could be a bug in virt-manager's assumptions of the build command, but even consider a virsh user: previously build just created a directory, now it formats a partition, without any XML change. Thanks, Cole

4 11

[libvirt] XenAPI remote storage - target path
by Sharadha Prabhakar (3P) 16 Mar '10

16 Mar '10

Hi, I'm trying to write a Remote Storage driver for XenAPI. I see that target-path is used for both storage pools and volumes. In the case of XenAPI remote storage, the storage is not mounted on the local host where libvirt is running. Storage is maintained in a remote location only. In this case how do I specify target-path and how do I go about creating a VM with storage using libvirt APIs with virsh and virt-manager. Virt-manager expects me to give an absolute path for target-path. But In my case I don't have an absolute path. Can I have a target path like this for a particular storage volume "/storage pool uuid/storage-vol uuid"? This will help me identify which storage pool is Libvirt talking about and which volume in it. Using this information I can fetch data from The remote location and give it back to libvirt. Is this approach ok or does libvirt support my specific remote storage case in some way. Could someone clarify please? Regards, Sharadha

2 2

[libvirt] RFC: Add further domain event callbacks
by Daniel P. Berrange 16 Mar '10

16 Mar '10

In the current domain APIs, we currently have support for getting notified of domain lifecycle transition events. THis is done using two methods int virConnectDomainEventRegister(virConnectPtr conn, virConnectDomainEventCallback cb, void *opaque, virFreeCallback freecb); int virConnectDomainEventDeregister(virConnectPtr conn, virConnectDomainEventCallback cb); This allows an app to register a callback that looks like this typedef int (*virConnectDomainEventCallback)(virConnectPtr conn, virDomainPtr dom, int event, int detail, void *opaque); Where 'event' is the lifecycle transition (suspended, stopped, started, etc) and 'detail' is the cause of he transition (pause, migration, shutdown, etc) I have outstanding feature requests to add a lot more event notifications to the libvirt API. In particular - IO Errors. Parameter 'alias' the name of the block device with the error - Reboot. No parameters You can argue whether this should be part of the lifecycle events. On the one hand it is not guest visible, because the guest does a internal machine reset without the host ever seeing a change. This is different from other cases where QEMU itself is stopping/starting. - Watchdog. Parameter 'action', saying what is going to happen to the guest due to this watchdog firing (ignored, shutdown, paused, etc) - VNC client. In fact three events, connect, authenticated and disconnect. Parameters, TCP address & port number of client, and also of the server. Optionally a SASL username and TLS certificate name of the authenticated user - Guest user. Logon/logoff events. This requires co-operation from a guest agent, so it may not actually be of scope of libvirt. - Disk 'high watermark'. Emitted when a QCow volume grows beyond a certain physical allocation. THis allows an app to enlarge the underlying storage holding the qcow volume before an 'out of space' occurs. Parameter is the disk alias name. So we can see there are events with a wide variety of parameters and we need to figure out how to represent this in the API. Option 1 -------- Follow the existing lifecycle event model. For each new event, add a virConnectXXXXXEventRegister & virConnectEventDeregister method, and typedef a new callback for them. eg typedef int (*virConnectDomainBlockIOEventCallback)(virConnectPtr conn, virDomainPtr dom, const char *diskname, const char *alias, void *opaque); int virConnectDomainBlockIOEventRegister(virConnectPtr conn, virConnectDomainBlockIOEventCallback cb, void *opaque, virFreeCallback freecb); int virConnectDomainBlockIOEventDeregister(virConnectPtr conn, virConnectDomainEventCallback cb); So we'll have 2 extra APIs for every event + a new typedef. We'll also need to add new APIs in src/conf/domain_event.h to cope with dispatch Option 2 -------- GLib/GObject take a very loosely typed approach to registering/unregistering events. The have a single pair of methods that work for any event & a generic callback signature, requiring application casts. typedef int (*virConnectEventCallback)(void *opaque); int virConnectEventRegister(virConnectPtr conn, const char *eventname, virConnectEventCallback cb, void *opaque, virFreeCallback freecb); int virCOnnectEventUnregister(virConnectPtr conn, int eventID); In this model, the register method returns a unique integer ID for the callback which can be used to unregister it. Application's using this will still need a strongly typed callback for receiving the event, but when calling virConnectEventRegister(), the would do an explicit 'bad' cast to 'virConnectEventCallback' Option 3 -------- A hybrid of both approaches. Have a new 'register' method for each type of event that takes a strongly typed callback, but have a generic 'unregister' method that just uses the 'int eventID' int virConnectDomainBlockIOEventRegister(virConnectPtr conn, virConnectDomainBlockIOEventCallback cb, void *opaque, virFreeCallback freecb); int virCOnnectEventUnregister(virConnectPtr conn, int eventID); Option 4 -------- Have one pair of register/unregister events, but instead of passing diffeerent parameters to each callback, have a generic callback that takes a single parameter. This parameter would be declared as a union. So depending on the type of event being received, you'd access different parts of the union typedef union { virConnectDomainBlockIOEvent blockio; virConnectDomainWatchdogEvent watchdog; ...other events... } virConnectEvent; Either we could include a dummy member in the union with padding to 1024 bytes in size for future expansion, or we could simply declare that apps must never allocate this data type themselves, thus allowing us to enlarge it at will. typedef int (*virConnectEventCallback)(int eventType, virConnectEvent, void *opaque); int virConnectEventRegister(virConnectPtr conn, const char *eventname, virConnectEventCallback cb, void *opaque, virFreeCallback freecb); int virConnectEventUnregister(virConnectPtr conn, int eventID); There is one final question unrelated to these 4 options. For the lifecycle events we always registered against the 'virConnectPtr' since that is needed to capture 'domain created' events where there's no virDomainPtr to register a callback against yet. Do we want to always register all events aganist the virConnectPtr, and then pass a 'virDomainPtr' as a parameter to the callbacks as needed. Or should we allow registering events against the virDomainPtr directly. The latter might make it simpler to map libvirt into GLib/GObjects event system in the future. Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://deltacloud.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

2 2

[libvirt] emulation failed (pagetable) rip
by Adam Mooz 15 Mar '10

15 Mar '10

Good morning list, I run a few VM's using libvirt on a C2Q 6600, have been steadily and happily since November, most of the VM's have been happily chugging away since then as well. I woke on Thursday to find one of the VM's had crashed, couldn't login and wasn't responding to the "virsh shutdown" commands. Now when I try to start it; it comes up as 'running' according to virsh but doesn't even request an IP from my DHCP server. I've enabled info logging and the log output looks like this: > Mar 14 09:52:55 Godfather kernel: [24816.021257] __ratelimit: 932597 callbacks suppressed > Mar 14 09:52:55 Godfather kernel: [24816.021262] emulation failed (pagetable) rip 7f2019b7aafb 66 0f 7f 07 > Mar 14 09:52:55 Godfather kernel: [24816.021297] emulation failed (pagetable) rip 7f2019b7aafb 66 0f 7f 07 > Mar 14 09:52:55 Godfather kernel: [24816.021321] emulation failed (pagetable) rip 7f2019b7aafb 66 0f 7f 07 > Mar 14 09:52:55 Godfather kernel: [24816.021344] emulation failed (pagetable) rip 7f2019b7aafb 66 0f 7f 07 Any ideas on fixing this? ----------------------------------------------------------------- Adam Mooz Adam.Mooz(a)gmail.com http://www.AdamMooz.com

2 2

[libvirt] [PATCH] security: Set permissions for kernel/initrd
by Cole Robinson 15 Mar '10

15 Mar '10

Fixes URL installs when running virt-install as root on Fedora. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/qemu/qemu_security_dac.c | 21 +++++++++++++++++++++ src/security/security_selinux.c | 16 ++++++++++++++++ 2 files changed, 37 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_security_dac.c b/src/qemu/qemu_security_dac.c index 6911f48..1883fbe 100644 --- a/src/qemu/qemu_security_dac.c +++ b/src/qemu/qemu_security_dac.c @@ -332,6 +332,15 @@ qemuSecurityDACRestoreSecurityAllLabel(virDomainObjPtr vm) vm->def->disks[i]) < 0) rc = -1; } + + if (vm->def->os.kernel && + qemuSecurityDACRestoreSecurityFileLabel(vm->def->os.kernel) < 0) + rc = -1; + + if (vm->def->os.initrd && + qemuSecurityDACRestoreSecurityFileLabel(vm->def->os.initrd) < 0) + rc = -1; + return rc; } @@ -356,6 +365,18 @@ qemuSecurityDACSetSecurityAllLabel(virDomainObjPtr vm) return -1; } + if (vm->def->os.kernel && + qemuSecurityDACSetOwnership(vm->def->os.kernel, + driver->user, + driver->group) < 0) + return -1; + + if (vm->def->os.initrd && + qemuSecurityDACSetOwnership(vm->def->os.initrd, + driver->user, + driver->group) < 0) + return -1; + return 0; } diff --git a/src/security/security_selinux.c b/src/security/security_selinux.c index b2c8581..975b315 100644 --- a/src/security/security_selinux.c +++ b/src/security/security_selinux.c @@ -616,6 +616,14 @@ SELinuxRestoreSecurityAllLabel(virDomainObjPtr vm) rc = -1; } + if (vm->def->os.kernel && + SELinuxRestoreSecurityFileLabel(vm->def->os.kernel) < 0) + rc = -1; + + if (vm->def->os.initrd && + SELinuxRestoreSecurityFileLabel(vm->def->os.initrd) < 0) + rc = -1; + return rc; } @@ -736,6 +744,14 @@ SELinuxSetSecurityAllLabel(virDomainObjPtr vm) return -1; } + if (vm->def->os.kernel && + SELinuxSetFilecon(vm->def->os.kernel, default_content_context) < 0) + return -1; + + if (vm->def->os.initrd && + SELinuxSetFilecon(vm->def->os.initrd, default_content_context) < 0) + return -1; + return 0; } -- 1.6.6.1

2 2

[libvirt] [PATCH] qemu: Fix USB by product with security enabled
by Cole Robinson 15 Mar '10

15 Mar '10

We need to call PrepareHostdevs to determine the USB device path before any security calls. PrepareHostUSBDevices was also incorrectly skipping all USB devices. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/qemu/qemu_driver.c | 11 ++++++----- 1 files changed, 6 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 26b5600..0fa9b2f 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2360,7 +2360,7 @@ qemuPrepareHostUSBDevices(struct qemud_driver *driver ATTRIBUTE_UNUSED, if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS) continue; - if (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI) + if (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB) continue; /* Resolve a vendor/product to bus/device */ @@ -2703,6 +2703,11 @@ static int qemudStartVMDaemon(virConnectPtr conn, return -1; } + DEBUG0("Preparing host devices"); + /* Must be run before security labelling */ + if (qemuPrepareHostDevices(driver, vm->def) < 0) + goto cleanup; + DEBUG0("Generating domain security label (if required)"); /* If you are using a SecurityDriver with dynamic labelling, then generate a security label for isolation */ @@ -2769,10 +2774,6 @@ static int qemudStartVMDaemon(virConnectPtr conn, if (qemuSetupCgroup(driver, vm) < 0) goto cleanup; - DEBUG0("Preparing host devices"); - if (qemuPrepareHostDevices(driver, vm->def) < 0) - goto cleanup; - if (VIR_ALLOC(priv->monConfig) < 0) { virReportOOMError(); goto cleanup; -- 1.6.6

3 3

[libvirt] compile libvirt in windows
by Luis Costa 15 Mar '10

15 Mar '10

Dear friends It would like to know as I I make to compile libvirt in the Windows? I already installed the MingGW, you I have some cake prescription? Regards, -- Luís Carlos Moreira da Costa Eclipse RAP, RCP, GMF, OSGI, Spring-DM and Pentaho Developer Regional Communities/Brazil http://wiki.eclipse.org/Regional_Communities/Brazil Java, write once, run everywhere™

2 1

[libvirt] libvirt modifying iptables rules
by Varrun Ramani 15 Mar '10

15 Mar '10

Hi I am right now undertaking a project which deals with verification of firewall rules. I wish to know which applications/libraries modify/query firewall rules. I came to know that libvirt modifies iptables rules. Can anyone let me know for what purposes/how libvirt modifies the rules? Thanks & Regards Varrun Ramani

4 3

[libvirt] [PATCH] qemu: pass the information when disks are read-only
by Daniel Veillard 15 Mar '10

15 Mar '10

The problem is that even when the user indicates that a drive has to be opened as read-only (for example a cd-rom) qemu didn't get the information and tried to open read-write. This could get really nasty in case SELinux is monitoring that no read-write access is attempted: https://bugzilla.redhat.com/show_bug.cgi?id=536760 Patch is relatively simple and based on teh assumption that if the -device qemu options are allowed then qemu also knows about the readonly option (this was added to the help string quite later so we should not rely on that in this case). * src/qemu/qemu_conf.c: add the ",readonly=on" for read-only disks and also parse it back in qemuParseCommandLineDisk() * tests/qemuxml2argvtest.c tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.args tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.xml: add a specific regression test diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 40ca221..fb23c52 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2398,6 +2398,9 @@ qemuBuildDriveStr(virDomainDiskDefPtr disk, if (bootable && disk->device == VIR_DOMAIN_DISK_DEVICE_DISK) virBufferAddLit(&opt, ",boot=on"); + if (disk->readonly && + qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) + virBufferAddLit(&opt, ",readonly=on"); if (disk->driverType && disk->type != VIR_DOMAIN_DISK_TYPE_DIR && qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE_FORMAT) @@ -4694,6 +4697,9 @@ qemuParseCommandLineDisk(const char *val, _("cannot parse drive unit '%s'"), val); goto cleanup; } + } else if (STREQ(keywords[i], "readonly")) { + if ((values[i] == NULL) || STREQ(values[i], "on")) + def->readonly = 1; } } diff --git a/tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.args b/tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.args new file mode 100644 index 0000000..fbc3226 --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.args @@ -0,0 +1 @@ +LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test /usr/bin/qemu -S -M pc -m 214 -smp 1 -nographic -nodefaults -monitor unix:/tmp/test-monitor,server,nowait -no-acpi -boot c -drive file=/dev/HostVG/QEMUGuest1,if=none,id=drive-ide0-0-0 -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0 -drive file=/dev/sr0,if=none,media=cdrom,id=drive-ide0-1-0,readonly=on -device ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0 -usb -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 diff --git a/tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.xml b/tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.xml new file mode 100644 index 0000000..39c3a1c --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-disk-drive-readonly-disk.xml @@ -0,0 +1,30 @@ +<domain type='qemu'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory>219200</memory> + <currentMemory>219200</currentMemory> + <vcpu>1</vcpu> + <os> + <type arch='i686' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <emulator>/usr/bin/qemu</emulator> + <disk type='block' device='disk'> + <source dev='/dev/HostVG/QEMUGuest1'/> + <target dev='hda' bus='ide'/> + <address type='drive' controller='0' bus='0' unit='0'/> + </disk> + <disk type='block' device='cdrom'> + <source dev='/dev/sr0'/> + <target dev='hdc' bus='ide'/> + <readonly/> + <address type='drive' controller='0' bus='1' unit='0'/> + </disk> + <controller type='ide' index='0'/> + </devices> +</domain> diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 4ca946a..e3762c9 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -249,6 +249,8 @@ mymain(int argc, char **argv) QEMUD_CMD_FLAG_DRIVE_BOOT | QEMUD_CMD_FLAG_DRIVE_FORMAT); DO_TEST("disk-drive-fat", QEMUD_CMD_FLAG_DRIVE | QEMUD_CMD_FLAG_DRIVE_BOOT | QEMUD_CMD_FLAG_DRIVE_FORMAT); + DO_TEST("disk-drive-readonly-disk", QEMUD_CMD_FLAG_DRIVE | + QEMUD_CMD_FLAG_DEVICE); DO_TEST("disk-drive-fmt-qcow", QEMUD_CMD_FLAG_DRIVE | QEMUD_CMD_FLAG_DRIVE_BOOT | QEMUD_CMD_FLAG_DRIVE_FORMAT); DO_TEST("disk-drive-shared", QEMUD_CMD_FLAG_DRIVE | -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

2 2

[libvirt] [PATCH] qemudDomainAttachSCSIDisk: handle empty controller list
by Jim Meyering 15 Mar '10

15 Mar '10

Clang found something that might be a real bug. I suspect that ...drive.controller will always be at least one, but we should not have to dive into the code trying to figure that out. It's easier/better here just to handle the potential trouble: clang saw that if it *was* zero, then the following "for" loop would not be entered, and "cont" would not be initialized. On the very next statement "cont" (uninitialized) would be dereferenced. >From 21ec1a8ae0218a5e8d789410318a973518ffec6c Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 2 Mar 2010 17:45:10 +0100 Subject: [PATCH] qemudDomainAttachSCSIDisk: handle empty controller list * src/qemu/qemu_driver.c (qemudDomainAttachSCSIDisk): Handle the (theoretical) case of an empty controller list, so that clang does not think the subsequent dereference of "cont" would dereference an undefined variable (due to preceding loop not iterating even once). --- src/qemu/qemu_driver.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 7f7c459..efb1857 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -5671,18 +5671,24 @@ static int qemudDomainAttachSCSIDisk(struct qemud_driver *driver, if (qemuAssignDeviceDiskAlias(disk, qemuCmdFlags) < 0) goto error; if (!(devstr = qemuBuildDriveDevStr(disk))) goto error; } if (!(drivestr = qemuBuildDriveStr(disk, 0, qemuCmdFlags))) goto error; + if (disk->info.addr.drive.controller <= 0) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, + _("no drive controller for %s"), disk->dst); + goto error; + } + for (i = 0 ; i <= disk->info.addr.drive.controller ; i++) { cont = qemuDomainFindOrCreateSCSIDiskController(driver, vm, i, qemuCmdFlags); if (!cont) goto error; } if (cont->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) { qemuReportError(VIR_ERR_INTERNAL_ERROR, _("SCSI controller %d was missing its PCI address"), cont->idx); -- 1.7.0.1.414.g89213d

4 6

[libvirt] eject ISO from cdrom
by Shi Jin 15 Mar '10

15 Mar '10

Hi there, I am using libvirt+qemu/kvm. For a guest VM that is running, I could change its cdrom media by running the following virsh command: virsh attach-disk <vm-id> <path-to-iso-file> hdc --driver file --type cdrom --mode readonly This will change the media in the cdrom for the guest VM. Once I am done, I can use virt-manager to disconnect the iso file to the cdrom. The resulting xml file looks like (only showing the relevant part) <disk type='file' device='cdrom'> <target dev='hdc' bus='ide'/> <readonly/> </disk> However, I am having trouble to do the same task (eject cd media) using virsh command. I am successful in running the "virsh attach-device 1 emptyCD.xml" command where the emptyCD.xml is an xml without the source file path, just like the one above. However, I want to do it by a single command without the need to create xml files. It should be something like > virsh attach-disk 1 "" hdc --driver file --type cdrom --mode readonly However, virsh does not like the empty string "". I also looked at the virt-manager source code, in this case, the python script is 912 if curpath: 913 # Disconnect cdrom 914 self.change_cdrom_media(dev_id_info, None, _type=None) Basically a None parameter is pass to the API. All I need to find out is how to do the same thing in the "virsh attach-disk" command. Do you have any suggestions/comments? Thanks a lot. Shi -- Shi Jin, PhD

2 1