November 2012 - Devel - libvirt List Archives

[libvirt] [PATCH 0/3] IPv6 enhancements; put dnsmasq parameters in conf-file
by Gene Czarcinski 29 Nov '12

29 Nov '12

These three patch files are packaged together because they serially depend on each other. These files have been rebased to "v3" of the dnsmasq capabilities and bind-dynamic patches. The DHCPv6 support checks dnsmasq's version and requires a minimum of 2.64. Also, using dnsmasq for providing the RA service is checked against the dnsmasq version and is currently 2.64. As with IPv4, IPv6 DHCP is only one subnetwork on an interface. Additionally, if other IPv6 addresses are defined, a warning message is issued since the Router Advertisement service will support only state-full (DHCP) or state-less (SLAAC) addressing on a network interface (not both). Thus, the additional subnetworks will need to be manually configured to properly function. If dnsmasq provides the RA service, it also points to itself as a RDNSS (Recursive DNS Server) as part of the information is supplies. If IPv6 DHCP is not being run, then SLAAC addressing is supported for any IPv6 addresses specified. Gene Czarcinski (3): v1: allow guest to guest IPv6 without gateway definition v8.1 add support for DHCPv6 v7.9: put dnsmasq parameters into conf-file docs/formatnetwork.html.in | 126 ++++- docs/schemas/network.rng | 12 +- src/conf/network_conf.c | 100 ++-- src/network/bridge_driver.c | 590 ++++++++++++++------- src/network/bridge_driver.h | 7 +- src/util/dnsmasq.c | 9 +- tests/networkxml2argvdata/dhcp6-nat-network.argv | 14 + tests/networkxml2argvdata/dhcp6-nat-network.xml | 24 + tests/networkxml2argvdata/dhcp6-network.argv | 14 + tests/networkxml2argvdata/dhcp6-network.xml | 14 + .../dhcp6host-routed-network.argv | 12 + .../dhcp6host-routed-network.xml | 19 + tests/networkxml2argvdata/isolated-network.argv | 25 +- .../networkxml2argvdata/nat-network-dns-hosts.argv | 14 +- .../nat-network-dns-srv-record-minimal.argv | 34 +- .../nat-network-dns-srv-record.argv | 24 +- .../nat-network-dns-txt-record.argv | 22 +- tests/networkxml2argvdata/nat-network.argv | 22 +- tests/networkxml2argvdata/netboot-network.argv | 28 +- .../networkxml2argvdata/netboot-proxy-network.argv | 26 +- tests/networkxml2argvdata/routed-network.argv | 11 +- tests/networkxml2argvtest.c | 47 +- 22 files changed, 838 insertions(+), 356 deletions(-) create mode 100644 tests/networkxml2argvdata/dhcp6-nat-network.argv create mode 100644 tests/networkxml2argvdata/dhcp6-nat-network.xml create mode 100644 tests/networkxml2argvdata/dhcp6-network.argv create mode 100644 tests/networkxml2argvdata/dhcp6-network.xml create mode 100644 tests/networkxml2argvdata/dhcp6host-routed-network.argv create mode 100644 tests/networkxml2argvdata/dhcp6host-routed-network.xml -- 1.7.11.7

1 3

[libvirt] [PATCH 0/6] More coverity issues
by Ján Tomko 29 Nov '12

29 Nov '12

REVERSE_INULL, SIZEOF_MISMATCH and some uninitialized variables. Ján Tomko (6): conf: fix NULL check in virNetDevBandwidthParse util: fix virBitmap allocation in virProcessInfoGetAffinity virsh: use correct sizeof when allocating cpumap virsh: do timing even for unusable connections rpc: don't destroy xdr before creating it in virNetMessageEncodeHeader conf: fix uninitialized variable in virDomainListSnapshots src/conf/netdev_bandwidth_conf.c | 4 +++- src/conf/snapshot_conf.c | 2 +- src/rpc/virnetmessage.c | 2 +- src/util/processinfo.c | 2 +- tools/virsh-domain.c | 4 ++-- tools/virsh.c | 12 ++++++------ 6 files changed, 14 insertions(+), 12 deletions(-) -- 1.7.8.6

2 13

[libvirt] libvirt-tck breackage
by Guido Günther 29 Nov '12

29 Nov '12

Hi, it seems a recent commit broke allmost all off the libvirt-tck tests: http://honk.sigxcpu.org:8001/job/libvirt-tck-debian-wheezy-qemu-session/196… I won't have a chance to have a closer look but but wanted to at least send a heads up. I'll have a closer look later this week. Cheers, -- Guido

2 3

[libvirt] Globally Reserve Resources for Host
by Dusty Mabe 29 Nov '12

29 Nov '12

Hi, I am interested in the capability to globally reserve resources(cpu and memory) for a KVM host. I know you can configure memory limits for each guest (http://libvirt.org/formatdomain.html#elementsMemoryTuning) but would like the ability to reserve host cpu and memory without having to actively do it by modifying each guests xml. For clarity, what I mean by "reserve resources" is that there are certain cpus and a certain amount of memory that guests will never have access to. This can be achieved using cgroups. Does anyone think this functionality would be useful? This is primarily to prevent the host from being starved when the allocation of guests have the host overcommitted/oversubscribed. Note: I think vmware has a similar functionality, but I am not sure as I don't really use vmware http://blogs.technet.com/b/virtualpfe/archive/2011/08/29/hyper-v-dynamic-me… Thanks for any thoughts, Dusty

3 6

[libvirt] [PATCH v2 0/4] Support for sending signals to guest processes
by Daniel P. Berrange 29 Nov '12

29 Nov '12

A follow up to https://www.redhat.com/archives/libvir-list/2012-November/msg01283.html Changed in this version - Switch pid to long long to allow process groups & 64-bit pids - Add explicit enum entries for RT signals - Do case insensitive parsing in virsh - Fix version numbers - Improve docs

2 10

[libvirt] [PATCH] network: fix crash when portgroup has no name
by Laine Stump 29 Nov '12

29 Nov '12

This resolves: https://bugzilla.redhat.com/show_bug.cgi?id=879473 The name attribute is required for portgroup elements (yes, the RNG specifies that), and there is code in libvirt that assumes it is non-null. Unfortunately, the portgroup parsing function wasn't checking for lack of portgroup. One adverse result of this was that attempts to update a network by adding a portgroup with no name would cause libvirtd to segfault. For example: virsh net-update default add portgroup "<portgroup default='yes'/>" This patch causes virNetworkPortGroupParseXML to fail if no name is specified, thus avoiding any later problems. --- src/conf/network_conf.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index 228951d..6ce2e63 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -1175,6 +1175,12 @@ virNetworkPortGroupParseXML(virPortGroupDefPtr def, /* grab raw data from XML */ def->name = virXPathString("string(./@name)", ctxt); + if (!def->name) { + virReportError(VIR_ERR_XML_ERROR, "%s", + _("Missing required name attribute in portgroup")); + goto error; + } + isDefault = virXPathString("string(./@default)", ctxt); def->isDefault = isDefault && STRCASEEQ(isDefault, "yes"); -- 1.7.11.7

2 3

[libvirt] [PATCH] qemu: Remove full stop from error messages
by Jiri Denemark 29 Nov '12

29 Nov '12

--- Pushed as trivial. src/qemu/qemu_agent.c | 2 +- src/qemu/qemu_monitor.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_agent.c b/src/qemu/qemu_agent.c index 7062d53..893f7f2 100644 --- a/src/qemu/qemu_agent.c +++ b/src/qemu/qemu_agent.c @@ -242,7 +242,7 @@ qemuAgentOpenUnix(const char *monitor, pid_t cpid, bool *inProgress) if (ret != 0) { virReportSystemError(errno, "%s", - _("monitor socket did not show up.")); + _("monitor socket did not show up")); goto error; } diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index fe8424f..aef5044 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -297,7 +297,7 @@ qemuMonitorOpenUnix(const char *monitor, pid_t cpid) if (ret != 0) { virReportSystemError(errno, "%s", - _("monitor socket did not show up.")); + _("monitor socket did not show up")); goto error; } -- 1.8.0

1 0

[libvirt] [PATCH 00/10] Misc LXC related fixes
by Daniel P. Berrange 29 Nov '12

29 Nov '12

The following is a list of random bug fixes I've identified while working on LXC

2 17

[libvirt] [test-API][PATCH v3] Add test case of set vcpus with flags
by Wayne Sun 29 Nov '12

29 Nov '12

v2: break down the case to small cases with separate flags * Use setVcpusFlags API to set domain vcpus with flags * 3 cases added, each only deal with one set flag value as in config, live or maximum * cases are independent on domain states, API will report error if not suitable for certain states * the sample conf is only one scenario of hotplug domain vcpus v3: merge config and maximum case to config * maximum flag can only work when domain is shutoff, merge it to config case to simplify code Signed-off-by: Wayne Sun <gsun(a)redhat.com> --- cases/set_vcpus_flags.conf | 67 +++++++++++++++++++++++++ repos/setVcpus/set_vcpus_config.py | 93 ++++++++++++++++++++++++++++++++++ repos/setVcpus/set_vcpus_live.py | 96 ++++++++++++++++++++++++++++++++++++ 3 files changed, 256 insertions(+), 0 deletions(-) create mode 100644 cases/set_vcpus_flags.conf create mode 100644 repos/setVcpus/__init__.py create mode 100644 repos/setVcpus/set_vcpus_config.py create mode 100644 repos/setVcpus/set_vcpus_live.py diff --git a/cases/set_vcpus_flags.conf b/cases/set_vcpus_flags.conf new file mode 100644 index 0000000..6cf595f --- /dev/null +++ b/cases/set_vcpus_flags.conf @@ -0,0 +1,67 @@ +domain:install_linux_cdrom + guestname + $defaultname + guestos + $defaultos + guestarch + $defaultarch + vcpu + $defaultvcpu + memory + $defaultmem + hddriver + $defaulthd + nicdriver + $defaultnic + imageformat + qcow2 + +domain:destroy + guestname + $defaultname + +setVcpus:set_vcpus_config + guestname + $defaultname + vcpu + 1 + maxvcpu + 8 + +domain:start + guestname + $defaultname + +setVcpus:set_vcpus_live + guestname + $defaultname + vcpu + 3 + username + $username + password + $password + +setVcpus:set_vcpus_config + guestname + $defaultname + vcpu + 5 + +domain:destroy + guestname + $defaultname + +domain:start + guestname + $defaultname + +domain:destroy + guestname + $defaultname + +domain:undefine + guestname + $defaultname + +options cleanup=enable diff --git a/repos/setVcpus/__init__.py b/repos/setVcpus/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/repos/setVcpus/set_vcpus_config.py b/repos/setVcpus/set_vcpus_config.py new file mode 100644 index 0000000..08eb53f --- /dev/null +++ b/repos/setVcpus/set_vcpus_config.py @@ -0,0 +1,93 @@ +#!/usr/bin/env python +# Test set domain vcpu with flag VIR_DOMAIN_AFFECT_CONFIG, also set +# and check max vcpu with flag VIR_DOMAIN_VCPU_MAXIMUM if maxvcpu +# param is given + +from xml.dom import minidom + +import libvirt +from libvirt import libvirtError + +from src import sharedmod + +required_params = ('guestname', 'vcpu', ) +optional_params = {'maxvcpu': 8, + } + +def get_vcpu_number(domobj): + """dump domain config xml description to get vcpu number, return + current vcpu and maximum vcpu number + """ + try: + guestxml = domobj.XMLDesc(2) + logger.debug("domain %s xml is :\n%s" %(domobj.name(), guestxml)) + xml = minidom.parseString(guestxml) + vcpu = xml.getElementsByTagName('vcpu')[0] + maxvcpu = int(vcpu.childNodes[0].data) + logger.info("domain max vcpu number is: %s" % maxvcpu) + + if vcpu.hasAttribute('current'): + attr = vcpu.getAttributeNode('current') + current = int(attr.nodeValue) + else: + logger.info("no 'current' atrribute for element vcpu") + current = int(vcpu.childNodes[0].data) + + logger.info("domain current vcpu number is: %s" % current) + + except libvirtError, e: + logger.error("libvirt call failed: " + str(e)) + return False + + return current, maxvcpu + +def set_vcpus_config(params): + """set domain vcpu with config flag and check, also set and check + max vcpu with maximum flag if optional param maxvcpu is given + """ + global logger + logger = params['logger'] + params.pop('logger') + guestname = params['guestname'] + vcpu = int(params['vcpu']) + maxvcpu = params.get('maxvcpu', None) + + logger.info("the name of virtual machine is %s" % guestname) + logger.info("the given vcpu number is %s" % vcpu) + + conn = sharedmod.libvirtobj['conn'] + + try: + domobj = conn.lookupByName(guestname) + logger.info("set domain vcpu as %s with flag: %s" % + (vcpu, libvirt.VIR_DOMAIN_AFFECT_CONFIG)) + domobj.setVcpusFlags(vcpu, libvirt.VIR_DOMAIN_AFFECT_CONFIG) + logger.info("set domain vcpu succeed") + + if maxvcpu: + logger.info("the given max vcpu number is %s" % maxvcpu) + logger.info("set domain maximum vcpu as %s with flag: %s" % + (maxvcpu, libvirt.VIR_DOMAIN_VCPU_MAXIMUM)) + domobj.setVcpusFlags(int(maxvcpu), libvirt.VIR_DOMAIN_VCPU_MAXIMUM) + logger.info("set domain vcpu succeed") + + except libvirtError, e: + logger.error("libvirt call failed: " + str(e)) + return 1 + + logger.info("check domain config xml to get vcpu number") + ret = get_vcpu_number(domobj) + if ret[0] == vcpu: + logger.info("domain current vcpu is equal as set") + if maxvcpu: + if ret[1] == int(maxvcpu): + logger.info("domain max vcpu is equal as set") + return 0 + else: + logger.error("domain max vcpu is not equal as set") + return 1 + else: + return 0 + else: + logger.error("domain current vcpu is not equal as set") + return 1 diff --git a/repos/setVcpus/set_vcpus_live.py b/repos/setVcpus/set_vcpus_live.py new file mode 100644 index 0000000..35a2976 --- /dev/null +++ b/repos/setVcpus/set_vcpus_live.py @@ -0,0 +1,96 @@ +#!/usr/bin/env python +# Test set domain vcpu with flag VIR_DOMAIN_VCPU_LIVE. Check +# domain xml and inside domain to get current vcpu number. The +# live flag only work on running domain, so test on shutoff +# domain will fail. + +from xml.dom import minidom + +import libvirt +from libvirt import libvirtError + +from src import sharedmod +from utils import utils + +required_params = ('guestname', 'vcpu', 'username', 'password', ) +optional_params = {} + +def get_current_vcpu(domobj, username, password): + """dump domain live xml description to get current vcpu number + and check in domain to confirm + """ + try: + guestxml = domobj.XMLDesc(1) + guestname = domobj.name() + logger.debug("domain %s xml is :\n%s" %(guestname, guestxml)) + xml = minidom.parseString(guestxml) + vcpu = xml.getElementsByTagName('vcpu')[0] + + if vcpu.hasAttribute('current'): + attr = vcpu.getAttributeNode('current') + current = int(attr.nodeValue) + else: + logger.info("no 'current' atrribute for element vcpu") + current = int(vcpu.childNodes[0].data) + + logger.info("domain current vcpu number in live xml is: %s" % current) + + except libvirtError, e: + logger.error("libvirt call failed: " + str(e)) + return False + + logger.debug("get the mac address of vm %s" % guestname) + mac = utils.get_dom_mac_addr(guestname) + logger.debug("the mac address of vm %s is %s" % (guestname, mac)) + + logger.info("check cpu number in domain") + ip = utils.mac_to_ip(mac, 180) + + cmd = "cat /proc/cpuinfo | grep processor | wc -l" + ret, output = utils.remote_exec_pexpect(ip, username, password, cmd) + if not ret: + logger.info("cpu number in domain is %s" % output) + if int(output) == current: + logger.info("cpu in domain is equal to current vcpu value") + else: + logger.error("current vcpu is not equal as check in domain") + return False + else: + logger.error("check in domain fail") + return False + + return current + +def set_vcpus_live(params): + """set domain vcpu with live flag and check + """ + global logger + logger = params['logger'] + params.pop('logger') + guestname = params['guestname'] + vcpu = int(params['vcpu']) + username = params['username'] + password = params['password'] + + logger.info("the name of virtual machine is %s" % guestname) + logger.info("the given vcpu number is %s" % vcpu) + + conn = sharedmod.libvirtobj['conn'] + + try: + domobj = conn.lookupByName(guestname) + logger.info("set domain vcpu as %s with flag: %s" % + (vcpu, libvirt.VIR_DOMAIN_VCPU_LIVE)) + domobj.setVcpusFlags(vcpu, libvirt.VIR_DOMAIN_VCPU_LIVE) + except libvirtError, e: + logger.error("libvirt call failed: " + str(e)) + return 1 + + logger.info("check domain vcpu") + ret = get_current_vcpu(domobj, username, password) + if ret == vcpu: + logger.info("domain vcpu is equal as set") + return 0 + else: + logger.error("domain vcpu is not equal as set") + return 1 -- 1.7.1

2 1

[libvirt] [PATCH 0/3] qemu: QMP Capability Probing Fixes
by Viktor Mihajlovski 29 Nov '12

29 Nov '12

QMP Capability probing will fail if the QEMU process cannot create the monitor socket file in /var/lib/libvirt/qemu. This is the case if the configured QEMU user is not root, but QEMU is run under root to perfom the probing. The suggested solution is to run QEMU as qemu user for probing as well. As it happens, this developed into a mini-series: it was necessary to let libvirt handle the pid file as this is stored in root-owned directory /var/run/libvirt/qemu. This prompted a race condition opening a socket. Last but not least caps->version was not filled with QMP probing. Viktor Mihajlovski (3): qemu: Wait for monitor socket even without pid qemu: Fix QMP Capabability Probing Failure qemu: Add QEMU version computation to QMP probing src/qemu/qemu_capabilities.c | 89 ++++++++++++++++++++++++++++++++++---------- src/qemu/qemu_capabilities.h | 7 +++- src/qemu/qemu_driver.c | 4 +- src/qemu/qemu_monitor.c | 2 +- 4 files changed, 78 insertions(+), 24 deletions(-) -- 1.7.12.4

3 9