- Devel - libvirt List Archives

[libvirt] [PATCH] python: Better error reporting if 'import libvirtmod' fails
by Cole Robinson 19 May '09

19 May '09

Don't squash a possibly legitimate libvirtmod error (e.g. some from clashing libvirt.so versions) with 'Cannot import cygvirtmod' --- python/libvir.py | 8 ++++++-- 1 files changed, 6 insertions(+), 2 deletions(-) diff --git a/python/libvir.py b/python/libvir.py index 8a16dd0..3cda8dc 100644 --- a/python/libvir.py +++ b/python/libvir.py @@ -8,8 +8,12 @@ # On cygwin, the DLL is called cygvirtmod.dll try: import libvirtmod -except: - import cygvirtmod as libvirtmod +except ImportError, lib_e: + try: + import cygvirtmod as libvirtmod + except ImportError, cyg_e: + if str(cyg_e).count("No module named"): + raise lib_e import types -- 1.6.2.2

3 23

[libvirt] [PATCH v2 0/2] FS backend VolCreateXMLFrom implementation.
by Cole Robinson 19 May '09

19 May '09

The following 2 patches implement virStorageVolCreateXMLFrom for the storage driver and the FS storage backend. Changes from v1: - Mark origvol as building in driver implementation - Make sure driver is locked when relocking pool object - Simplify sparse file detection with memcmp - Allocate read buffer on the heap. Thanks, Cole Cole Robinson (2): Storage driver implementation for CreateXMLFrom VolumeCreateXMLFrom FS storage backend implementation. src/storage_backend.h | 2 + src/storage_backend_fs.c | 247 +++++++++++++++++++++++++++++++++++++++------- src/storage_driver.c | 159 +++++++++++++++++++++++++++++- 3 files changed, 369 insertions(+), 39 deletions(-)

3 6

[libvirt] libvirt API to write sectors on a device
by Francesco Latino 19 May '09

19 May '09

Hello,***I am able to compile libvirt samples on my KVM hypervisor. (I edited and recompiled info.c code) *I am creating a sample code to read/write disk devices. With method *virDomainBlockPeek I can** read sectors from a given device attached to a KVM virtual guest.**I am not able to* find an API to write sectors on a disk: am I missing something ?Is there such method to write sectors from a given virtual disk ? Thanks,Francesco

2 2

[libvirt] [PATCH 0/5] Interface Config Public APIs and remote stubs, with MAC addresses in string format
by Laine Stump 19 May '09

19 May '09

These patches are based on those I sent last week (the public virInterface*() API definition, the local plumbing, and the RPC glue), but with suggestions incorporated: 1) MAC address is always used in null-terminated ASCII string format. This eliminates any potential problems with extra long addresses. 2) no comments in libvirt.h 3) flags arguments are all unsigned. Item (1) caused changes in most of the patches, so I just regenerated the entire set. Is this the proper/expected thing to do, or should I be making patches of patches?

3 9

[libvirt] PATCH: Enable migration with QEMU >= 0.10.0
by Daniel P. Berrange 19 May '09

19 May '09

The KVM migration code was added to QEMU for the 0.10.0 release, so we should enable this in libvirt now. Daniel diff -r be7993675e07 src/qemu_conf.c --- a/src/qemu_conf.c Thu Apr 30 14:49:27 2009 +0100 +++ b/src/qemu_conf.c Thu Apr 30 15:08:45 2009 +0100 @@ -473,16 +473,13 @@ int qemudExtractVersionInfo(const char * /* * Handling of -incoming arg with varying features - * -incoming tcp (kvm >= 79) - * -incoming exec (kvm >= 80) + * -incoming tcp (kvm >= 79, qemu >= 0.10.0) + * -incoming exec (kvm >= 80, qemu >= 0.10.0) * -incoming stdio (all earlier kvm) * * NB, there was a pre-kvm-79 'tcp' support, but it * was broken, because it blocked the monitor console * while waiting for data, so pretend it doesn't exist - * - * XXX when next QEMU release after 0.9.1 arrives, - * we'll need to add MIGRATE_QEMU_TCP/EXEC here too */ if (kvm_version >= 79) { flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_TCP; @@ -490,6 +487,9 @@ int qemudExtractVersionInfo(const char * flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_EXEC; } else if (kvm_version > 0) { flags |= QEMUD_CMD_FLAG_MIGRATE_KVM_STDIO; + } else if (version >= 10000) { + flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_TCP; + flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_EXEC; } if (retversion) -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

3 8

[libvirt] PATCH: Declare QEMU migration capabilities
by Daniel P. Berrange 18 May '09

18 May '09

This patches adds the missing data in the capabilities XML about the QEMU driver's migration support. Specifically it declares support for both offline and live migrate, and registers the 'tcp' URI scheme. Daniel diff -r 00f2e113eca8 src/qemu_conf.c --- a/src/qemu_conf.c Mon May 18 12:38:18 2009 +0100 +++ b/src/qemu_conf.c Mon May 18 13:15:33 2009 +0100 @@ -371,7 +371,7 @@ virCapsPtr qemudCapsInit(void) { uname (&utsname); if ((caps = virCapabilitiesNew(utsname.machine, - 0, 0)) == NULL) + 1, 1)) == NULL) goto no_memory; /* Using KVM's mac prefix for QEMU too */ @@ -380,6 +380,9 @@ virCapsPtr qemudCapsInit(void) { if (virCapsInitNUMA(caps) < 0) goto no_memory; + virCapabilitiesAddHostMigrateTransport(caps, + "tcp"); + /* First the pure HVM guests */ for (i = 0 ; i < ARRAY_CARDINALITY(arch_info_hvm) ; i++) if (qemudCapsInitGuest(caps, -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

2 1

[libvirt] Domain Event Handlers
by Noah Zoschke 18 May '09

18 May '09

Hello, I have been playing around with the libvirt domain event handlers in Python and am seeing something surprising. I am using trunk libvirt and `event-test.py` with some more of the debugging strings uncommented. When I trigger my first event: virsh # define xppro-sp2-0.img.xml Domain xppro-sp2-0.img defined from xppro-sp2-0.img.xml The timeout and handle callbacks are called, but not my domainEventRegister handler: Invoking Timeout CB Invoking Handle CB When I trigger a second event: virsh # create xppro-sp2-0.img.xml Domain xppro-sp2-0.img created from xppro-sp2-0.img.xml my event handler for the first event is finally called: Invoking Timeout CB domainEventHandler EVENT: Domain xppro-sp2-0.img(-1) Added 1 Invoking Handle CB Likewise for a third event: virsh # destroy xppro-sp2-0.img Domain xppro-sp2-0.img destroyed Invoking Timeout CB domainEventHandler EVENT: Domain xppro-sp2-0.img(1) Started 0 Invoking Handle CB So my handler is always one event behind. This behavior is not intended is it? Am I missing either something in `event-test.py` or in how I am using virsh that would correct this behavior? Thank you, Noah

3 3

[libvirt] Usage of UUID
by Eduardo Otubo 18 May '09

18 May '09

Hello all, When I started to implement phyp driver, I use to handle the UUID as a normal 'unsigned char *' with a LPAR value called "logical_serial_num" that is a string in the format "067BE8B2", an identifier. In the last comments I got from danpb and DV, they called my attention to const string at UUID handling and some funtcions that must be implemented: domainDumpXML, listDefinedDomains and numOfDefinedDomains. My question is: Once I don't have information from my hypervisor about UUID, do I have to handle UUID at "list" or "list --all" commands? Or I should keep and handle this information only when defining a domaing at libvirt? I mean, in my point of view, there is no need to generate and store a UUID while just listing LPARs in the remote server. What I should do is to generate a random UUID (and store it) just when the user decides to define it at libvirt. Am I thinking right? Thanks for the help. []'s -- Eduardo Otubo Software Engineer Linux Technology Center IBM Systems & Technology Group Mobile: +55 19 8135 0885 otubo(a)linux.vnet.ibm.com

2 1

[libvirt] is qemu really needed for make rpm or is kvm sufficient?
by Gerrit Slomma 18 May '09

18 May '09

Hello I am curious if qemu is really needed to do "make rpm". As far as i have seen only the src.rpm has this dependency but neither of the install-rpms. What is this needed for in the src.rpm-package that do lack the install-packages? I haven't installed qemu on my hosts, only kvm and everything works fine, even live-migration works now without the extra qemu installed on the host i build the libvirt. But now "make rpm" fails due to the dependency for qemu. Couldn't either: A) the qemu-requirement be patched out of the requirements for the src.rpm (if possible) or B) the kvm-package has qemu and qemu-img added as "provides" (okay - this would be another mailing list) Gerrit

2 1

[libvirt] patch for virsh list
by Gerrit Slomma 16 May '09

16 May '09

Hello I thought it would be cool if virsh list behaves a bit more like virt-manager or xm list because as a system-administrator i want to know with one look what VMs are active, how much memory they commit and what their runtime is. I implemented this with the following patch which give such an sample output. What is your opinion and/or have i missed something? [root@rr016 ~]# virsh list --all Id Name Status Mem(kB) VCPUs Time(s) --------------------------------------------------------------- 1 rr019v3 laufend 1572864 2 15,3 2 rr019v4 laufend 1572864 2 11,7 - OpenSolaris08.11 ausschalten - Win7b_64 ausschalten --- a/src/virsh.c 2009-04-15 21:43:26.000000000 +0200 +++ b/src/virsh.c 2009-05-15 21:35:53.000000000 +0200 @@ -658,14 +658,16 @@ qsort(&names[0], maxname, sizeof(char*), namesorter); } } - vshPrintExtra(ctl, "%3s %-20s %s\n", _("Id"), _("Name"), _("State")); - vshPrintExtra(ctl, "----------------------------------\n"); + vshPrintExtra(ctl, "%3s %-20s %-15s %-8s %-5s %-7s\n", _("Id"), _("Name"), _("State"), + _("Mem(kB)"), _("VCPUs"), _("Time(s)")); + vshPrintExtra(ctl, "---------------------------------------------------------------\n"); for (i = 0; i < maxid; i++) { virDomainInfo info; virDomainPtr dom = virDomainLookupByID(ctl->conn, ids[i]); const char *state; - + double cpuUsed = 0; + /* this kind of work with domains is not atomic operation */ if (!dom) continue; @@ -675,10 +677,19 @@ else state = N_(vshDomainStateToString(info.state)); - vshPrint(ctl, "%3d %-20s %s\n", + if (info.cpuTime != 0) { + cpuUsed = info.cpuTime; + + cpuUsed /= 1000000000.0; + } + + vshPrint(ctl, "%3d %-20s %-15s %-8d %-5d %7.1f\n", virDomainGetID(dom), virDomainGetName(dom), - state); + state, + virDomainGetMaxMemory(dom), + info.nrVirtCpu, + cpuUsed); virDomainFree(dom); } for (i = 0; i < maxname; i++) {

2 2