- Devel - libvirt List Archives

[libvirt] Xen config files again
by Thomas Moyer 06 Jul '08

06 Jul '08

I've seen in the src tree that there are functions to read and use Xen configuration files. Are these functions available in the Python bindings or do I need to use C to get access to them? What I am mostly interested in doing is parsing the Xen config file and getting back the XML version that I can store and use later on. -- Thomas Moyer Graduate Student CSE Department Penn State University tmmoyer(a)cse.psu.edu http://www.cse.psu.edu/~tmmoyer

2 3

[libvirt] 'starting' a domain using Python libvirt..
by Michael March 06 Jul '08

06 Jul '08

I'm trying to figure out how to start a domain .. and I can't figure out what method to use.. When I do this: >>> dir(conn.lookupByName('my-vm-1')) I get these options: ['ID', 'OSType', 'UUID', 'UUIDString', 'XMLDesc', '__del__', '__doc__', '__init__', '__module__', '_conn', '_o', 'attachDevice', 'autostart', 'blockStats', 'connect', 'coreDump', 'create', 'destroy', 'detachDevice', 'info', 'interfaceStats', 'maxMemory', 'maxVcpus', 'migrate', 'name', 'pinVcpu', 'reboot', 'resume', 'save', 'setAutostart', 'setMaxMemory', 'setMemory', 'setVcpus', 'shutdown', 'suspend', 'undefine'] I don't see a method to start a domain up.. What am I missing here? thanks!

2 2

[libvirt] PATCH: 0/2: port Test driver to new XML apis
by Daniel P. Berrange 04 Jul '08

04 Jul '08

The next two patches port the test driver over to use the new XML APIs for dealing with domain and network XML. More details with each one... Daniel. -- |: 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 13

[libvirt] storage_conf.c: In function 'virStorageSize'
by Jaime Traver 04 Jul '08

04 Jul '08

Hi, I am having problem with compilation of libvirt, i has saw one post about this in the maillist, but it don't resolved the problem. The detail of erros is: storage_conf.c: In function 'virStorageSize': storage_conf.c:627: error: 'ULLONG_MAX' undeclared (first use in this function) storage_conf.c:627: error: (Each undeclared identifier is reported only once storage_conf.c:627: error: for each function it appears in.) make[2]: *** [libvirt_la-storage_conf.lo] Error 1 make[2]: se sale del directorio `/home/revartj/libvirt/libvirt-0.4.4/src' make[1]: *** [all-recursive] Error 1 make[1]: se sale del directorio `/home/revartj/libvirt/libvirt-0.4.4' make: *** [all] Error 2 Any information about SO: EREBO:/home/revartj/libvirt/libvirt-0.4.4# *uname -a* Linux EREBO 2.6.18-6-xen-686 #1 SMP Sat May 24 14:16:03 UTC 2008 i686 GNU/Linux EREBO:/home/revartj/libvirt/libvirt-0.4.4# *gcc -v* Using built-in specs. Target: i486-linux-gnu Configured with: ../src/configure -v --enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr --enable-shared --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --enable-nls --program-suffix=-4.1 --enable-__cxa_atexit --enable-clocale=gnu --enable-libstdcxx-debug --enable-mpfr --with-tune=i686 --enable-checking=release i486-linux-gnu Thread model: posix gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) Someboy know how i can resolved this error. Thank everybody by advance. revartj

2 1

[libvirt] beta libvirt-java port uploaded
by Daniel Schwager 03 Jul '08

03 Jul '08

Hi, we are using libvirt in our softwaredemo.de plattform (launching in 2 or 3 weeks).. Because I saw some thread relating to the java-port, I uploaded our modified version (SNAPSHOT of my eclipse-workspace, based on the Tóth István libvir-java port - e.g. http://www.mail-archive.com/libvir-list@redhat.com/msg03131.html) to http://www.softwaredemo.de/index.php?id=154 for free access. At the moment, it's does not support all the XML-elements, because if haven't enough time to update it out XSD-file - but I works with the current libvirt 0.4.4-1 (-: - and work for us .... The uploaded java-project depends on some other (internal) staff of our company (devtools, ivy-repository), so it will not compile out-of-the-box for you - sorry for this at the moment. The TGZ includes all Dependant libs in lib/compiler Dependant infos in output/ivy-report Output artifacts in output/artifact (jar-files for using to access libvirt ...) If there are people wanting to work-on/update/extend these sources, I will think about adapting our individual-company-dependant project-build-framework to run everywhere (replace our ivy-repos with public maven-repos, insert the missing ant-scripts, ...) Maybe it could be also a good input for other people's to restart the libvirt-java project for their own (look into the project and uses some fragemets/ideas - staf, testing, buildprocess, ..) Thx to libvirt-community Danny (from softwaredemo.de ;-) If you have question, please ask (-:

2 3

[libvirt] PATCH: Fix memory leak in OOM cleanup for capabilities
by Daniel P. Berrange 03 Jul '08

03 Jul '08

The combo of the OOM testing and valgrind identified a place where we fail to correctly free memory when cleaning up from an OOM failure in the NUMA code for capabilities. This patch addresses that problem Daniel diff -r 0f3ebb9e7958 src/capabilities.c --- a/src/capabilities.c Thu Jul 03 11:42:29 2008 +0100 +++ b/src/capabilities.c Thu Jul 03 14:24:20 2008 +0100 @@ -224,17 +224,19 @@ if (VIR_ALLOC(cell) < 0) return -1; - caps->host.numaCell[caps->host.nnumaCell] = cell; - if (VIR_ALLOC_N(caps->host.numaCell[caps->host.nnumaCell]->cpus, - ncpus) < 0) + if (VIR_ALLOC_N(cell->cpus, ncpus) < 0) { + VIR_FREE(cell); return -1; - memcpy(caps->host.numaCell[caps->host.nnumaCell]->cpus, + } + memcpy(cell->cpus, cpus, ncpus * sizeof(*cpus)); - caps->host.numaCell[caps->host.nnumaCell]->ncpus = ncpus; - caps->host.numaCell[caps->host.nnumaCell]->num = num; + cell->ncpus = ncpus; + cell->num = num; + + caps->host.numaCell[caps->host.nnumaCell] = cell; caps->host.nnumaCell++; return 0; -- |: 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] PATCH: Parallelize OOM testing
by Daniel P. Berrange 03 Jul '08

03 Jul '08

The OOM testing takes a long time to run. Add in valgrind, and it takes a very very very long time to run. My dev box has 8 cpus though and it'd be nice to use all 8 rather than just 1 to cut the time by a factor of 8. So this patch tweaks the control harness for the OOM testrig so that it can fork off multiple worker processes. Each worker is assigned a number from 1 -> n, and will be responsible for test all allocations failures divisible by their number. By default it still runs single process, but if you set VIR_TEST_MP=1 then we use sysconf() to get the actual number of online CPUs and fork off one worker per CPU. eg, so to run the OOM testing with valgrind, parallelized $ VIR_TEST_MP=1 VIR_TEST_OOM=1 make valgrind Or to run a test directly $ VIR_TEST_MP=1 VIR_TEST_OOM=1 valgrind --leak-check=full ./xmconfigtest Daniel diff -r 3548706d639f tests/testutils.c --- a/tests/testutils.c Thu Jul 03 14:24:20 2008 +0100 +++ b/tests/testutils.c Thu Jul 03 14:24:40 2008 +0100 @@ -330,7 +330,9 @@ int n; char *oomStr = NULL, *debugStr; int oomCount; - + int mp = 0; + pid_t *workers; + int worker = 0; if ((debugStr = getenv("VIR_TEST_DEBUG")) != NULL) { if (virStrToLong_ui(debugStr, NULL, 10, &testDebug) < 0) testDebug = 0; @@ -344,6 +346,13 @@ oomCount = 0; if (oomCount) testOOM = 1; + } + + if (getenv("VIR_TEST_MP") != NULL) { + mp = sysconf(_SC_NPROCESSORS_ONLN); + fprintf(stderr, "Using %d worker processes\n", mp); + if (VIR_ALLOC_N(workers, mp) < 0) + return EXIT_FAILURE; } if (testOOM) @@ -371,11 +380,27 @@ else fprintf(stderr, "%d) OOM of %d allocs ", testCounter, approxAlloc); + if (mp) { + int i; + for (i = 0 ; i < mp ; i++) { + workers[i] = fork(); + if (workers[i] == 0) { + worker = i + 1; + break; + } + } + } + /* Run once for each alloc, failing a different one and validating that the test case failed */ - for (n = 0; n < approxAlloc ; n++) { + for (n = 0; n < approxAlloc && (!mp || worker) ; n++) { + if ((n % mp) != (worker - 1)) + continue; if (!testDebug) { - fprintf(stderr, "."); + if (mp) + fprintf(stderr, "%d", worker); + else + fprintf(stderr, "."); fflush(stderr); } virAllocTestOOM(n+1, oomCount); @@ -383,6 +408,20 @@ if (((func)(argc, argv)) != EXIT_FAILURE) { ret = EXIT_FAILURE; break; + } + } + + if (mp) { + if (worker) { + _exit(ret); + } else { + int i, status; + for (i = 0 ; i < mp ; i++) { + waitpid(workers[i], &status, 0); + if (WEXITSTATUS(status) != EXIT_SUCCESS) + ret = EXIT_FAILURE; + } + VIR_FREE(workers); } } -- |: 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] <script path='...'/> ignored for interface types 'bridge' and 'network'
by Emele, Joshua M 03 Jul '08

03 Jul '08

Hello, I'm new to libvirt, but I've looked around the internet for answers to my question with little success. I'd like to have a configuration script do host network integration when starting a virtual machine using virsh. The documentation (http://libvirt.org/formatdomain.html#elementsNICSBridge) suggests that the interface configuration can include a script tag that describes the path to the configuration script. My domain configuration (/etc/libvirt/qemu/test1.xml) includes: <interface type='bridge'> <source bridge='br0'/> <script path='/etc/libvirt/qemu/test1-networking'/> </interface> Unfortunately this value is never used. I'm using: [root@thoroughbred qemu]# rpm -qi libvirt Name : libvirt Relocations: (not relocatable) Version : 0.4.3 Vendor: Fedora Project Release : 1.fc8 Build Date: Fri 13 Jun 2008 03:15:16 AM PDT Looking at the source, the script value is used when 'ethernet' is specified (libvirt-0.4.3/src/qemu_conf.c:2690): case QEMUD_NET_ETHERNET: { char arg[PATH_MAX]; if (snprintf(arg, PATH_MAX-1, "tap,ifname=%s,script=%s,vlan=%d", net->dst.ethernet.ifname, net->dst.ethernet.script, vlan) >= (PATH_MAX-1)) goto error; ADD_ARG_LIT(arg); } However, the script value is not used when 'bridge' or 'network' is specified (libvirt-0.4.3/src/qemu_conf.c:2313): snprintf(tapfdstr, sizeof(tapfdstr), "tap,fd=%d,script=,vlan=%d,ifname=%s", tapfd, vlan, ifname); Is there a way to force libvirt to use the specified script with interface type 'network' or 'bridge'? If not, are there plans in the future to do so? Cheers, Joshua Emele

2 1

[libvirt] [PATCH] Return VIR_ERR_NO_SUPPORT in qemu SetMemory()/SetVcpus() if domain is active.
by Kaitlin Rupert 02 Jul '08

02 Jul '08

qemu doesn't support setting the memory or vcpus of an active guest. In this case, use the VIR_ERR_NO_SUPPORT return code to indicate that the action failed because its not supported. Index: src/qemu_driver.c =================================================================== RCS file: /data/cvs/libvirt/src/qemu_driver.c,v retrieving revision 1.85 diff -u -p -r1.85 qemu_driver.c --- src/qemu_driver.c 10 Jun 2008 10:43:28 -0000 1.85 +++ src/qemu_driver.c 11 Jun 2008 20:35:09 -0000 @@ -2148,7 +2148,7 @@ static int qemudDomainSetMemory(virDomai } if (qemudIsActiveVM(vm)) { - qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR, + qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT, "%s", _("cannot set memory of an active domain")); return -1; } @@ -2404,7 +2404,7 @@ static int qemudDomainSetVcpus(virDomain } if (qemudIsActiveVM(vm)) { - qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR, "%s", + qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT, "%s", _("cannot change vcpu count of an active domain")); return -1; } -- Kaitlin Rupert IBM Linux Technology Center kaitlin(a)linux.vnet.ibm.com

3 5

[Libvir] Re: libvirt on mingw
by Richard W.M. Jones 02 Jul '08

02 Jul '08

Brecht Sanders wrote: > Hi, > I saw on the following link: > http://www.mail-archive.com/libvir-list@redhat.com/msg04103.html > that you are also trying to compile libvirt on win32. > I'm also attempting to do this, and I guess I got stuck at the same > point your post was about. > Have you in the mean time found an XDR implementation that compiles on > MinGW and that implements xdr_u_quad_t? > If you did, can you please tell me where to find it? You'll find the answer to this question and more if you look through the libvir-list archives for the current month: https://www.redhat.com/archives/libvir-list/2008-January/thread.html Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 12