December 2015 - Devel - libvirt List Archives

[libvirt] [PATCH 0/7] tests: Always fake root directory
by Andrea Bolognani 10 Dec '15

10 Dec '15

When mocking filesystem access in the test suite, we have always assumed it to be some subdirectory of /sys depending on the needs of the specific test case. This limits our flexibility, and will be a problem once we need to start mocking eg. /dev as well, or simply two different parts of the /sys filesystem at the same time[1]. Solve the issue by always using the temporary directory as the root directory for the mocked filesystem. The series is organized as follows: 1-2: Tiny cleanups 3-4: Change the mock libraries so that they build the proper directory structure 5: Change the name of the environment variable used to pass the temporary directory name to the mock libraries 6: Make it so we'll be able to mock different parts of the /sys filesystem at the same time 7: Update scsihosttest to use the same directory structure Cheers. [1] https://www.redhat.com/archives/libvir-list/2015-November/msg00532.html is an example of why we would need to do that Andrea Bolognani (7): tests: scsihost: Don't set LIBVIRT_FAKE_SYSFS_DIR tests: pcimock: Remove check for fakesysfsdir tests: pcimock: Use the temporary directory as fake root tests: cgroupmock: Use the temporary directory as fake root tests: Rename LIBVIRT_FAKE_SYSFS_DIR to LIBVIRT_FAKE_ROOT_DIR tests: Use more specific names for variables tests: scsihost: Use fakerootdir instead of fakesysfsdir tests/scsihosttest.c | 17 ++++++--- tests/vircgroupmock.c | 99 ++++++++++++++++++++++++++++---------------------- tests/vircgrouptest.c | 16 ++++---- tests/virhostdevtest.c | 16 ++++---- tests/virpcimock.c | 69 +++++++++++++++++++---------------- tests/virpcitest.c | 16 ++++---- 6 files changed, 129 insertions(+), 104 deletions(-) -- 2.5.0

2 10

[libvirt] [PATCH v2 0/8] Add perf and Intel CMT feature support
by Qiaowei Ren 09 Dec '15

09 Dec '15

The series mainly adds Intel CMT feature support into libvirt. CMT is new introduced PQos (Platform Qos) feature to monitor the usage of cache by applications running on the platform. Currently CMT patches has been merged into Linux kernel mainline. The CMT implementation in Linux kernel is based on perf mechanism and there is no support for perf in libvirt, and so this series firstly add perf support into libvirt, including two public API and a set of util interfaces. And based on these APIs and interfaces, thie series implements CMT perf event support. TODO: 1. add support for new APIs into libvirt-python library. Changes since v1: * change perf APIs implementation to match new style of the API. * add new xml element * reenable all perf events previously enabled when libvirtd daemon restart. * redesign perf util functions. Qiaowei Ren (8): perf: add new public APIs for perf event perf: implement the remote protocol for perf event perf: implement a set of util functions for perf event qemu_driver: add support to perf event perf: add new xml element perf: reenable perf events when libvirtd restart virsh: implement new command to support perf virsh: extend domstats command daemon/remote.c | 47 ++++++ docs/schemas/domaincommon.rng | 27 ++++ include/libvirt/libvirt-domain.h | 19 +++ include/libvirt/virterror.h | 1 + src/Makefile.am | 1 + src/conf/domain_conf.c | 37 +++++ src/conf/domain_conf.h | 10 ++ src/driver-hypervisor.h | 12 ++ src/libvirt-domain.c | 93 ++++++++++++ src/libvirt_private.syms | 12 ++ src/libvirt_public.syms | 6 + src/qemu/qemu_domain.h | 3 + src/qemu/qemu_driver.c | 206 +++++++++++++++++++++++++++ src/qemu/qemu_process.c | 9 ++ src/remote/remote_driver.c | 39 +++++ src/remote/remote_protocol.x | 30 +++- src/remote_protocol-structs | 18 +++ src/util/virerror.c | 1 + src/util/virperf.c | 298 +++++++++++++++++++++++++++++++++++++++ src/util/virperf.h | 61 ++++++++ tools/virsh-domain-monitor.c | 7 + tools/virsh-domain.c | 125 ++++++++++++++++ tools/virsh.pod | 25 +++- 23 files changed, 1084 insertions(+), 3 deletions(-) create mode 100644 src/util/virperf.c create mode 100644 src/util/virperf.h -- 1.9.1

3 18

[libvirt] FD passing for chardevs and chardev backend multiplexing
by Daniel P. Berrange 09 Dec '15

09 Dec '15

Historically libvirt has connected stdout & stderr from QEMU directly to a plain file (/var/log/libvirt/qemu/$GUESTNAME.log). This has worked well enough in general, but is susceptible to a guest denial of service if the guest can cause QEMU to spew messages to stderr. There are enough places in QEMU and SPICE that still print to stderr that this isn't very hard to achieve. So In libvirt 1.3.0 we introduce a new daemon 'virtlogd' which is used for handling log file writing. When this is used, QEMU's stdout/stderr will be connected to an anonymous pipe file descriptor, the other end of which is held by the virtlogd daemon. The virtlogd daemon will only permit a fixed file size to be created before rotating the log file, so we no longer have the possibility of unbounded disk usage, which is nice. I'm now looking to extend the use of 'virtlogd' to also handle character devices. OpenStack has historically configured the primary serial port to log to a file in order to capture kernel boot up messages and later report them to the user via its API. This serial port file backend is of course susceptible to the same disk space denial of service. We don't really want to push file rotation logic into QEMU because that would involve giving QEMU permission to create / rename files, which is undesirable from a security POV. We also prefer a solution that ideally works with existing QEMU builds. So for this my plan is to stop using the QEMU 'file' backend for char devs and instead pass across a pre-opened file descriptor, connected to virtlogd. There is no "officially documented" way to pass in a file descriptor to QEMU chardevs, but since QEMU uses qemu_open(), we can make use of the fdset feature to achieve this. eg eg, consider fd 33 is the write end of a pipe file descriptor I can (in theory) do -add-fd set=2,fd=33 -chardev file,id=charserial0,path=/dev/fdset/2 Now in practice this doesn't work, because qmp_chardev_open_file() passes the O_CREAT|O_TRUNC flags in, which means the qemu_open() call will fail when using the pipe FD pased in via fdsets. After more investigation I found it *is* possible to use a socketpair and a pipe backend though... -add-fd set=2,fd=33 -chardev pipe,id=charserial0,path=/dev/fdset/2 ..because for reasons I don't understand, if QEMU can't open $PATH.in and $PATH.out, it'll fallback to just opening $PATH in read-write mode even. AFAICT, this is pretty useless with pipes since they are unidirectional, but, it works nicely with socketpairs, where virtlogd has one of the socketpairs and QEMU gets passed the other via fdset. I can easily check this works for historical QEMU versions back to when fdsets support was added to chardevs, but I'm working if the QEMU maintainers consider this usage acceptable over the long term, and if so, should we explicitly document it as supported ? If not, should we introduce a more explicit syntax for passing in a pre-opened FD for chardevs ? eg -add-fd set=2,fd=33 -chardev fd,id=charserial0,path=/dev/fdset/2 Or just make -chardev file,id=charserial0,path=/dev/fdset/2 actually work ? Or something else ? OpenStack has a further requirement to allow use of the serial port as an interactive console, at the same time that it is logging to a file which is something QEMU can't support at all currently. This essentially means being able to have multiple chardev backends all connected to the same serial frontend - specifically we would need a TCP backend and a file backend concurrently. Again this could be implemented in QEMU, but we'd prefer something that works with existing QEMU. This is not too difficult to achieve with virtlogd really. Instead of using the QEMU 'tcp' or 'unix' chardev protocol, we'd just always pass QEMU a pre-opened socketpair, and then leave the TCP/UNIX socket listening to the virtlogd daemon. This is portable with existing QEMU versions, but the obvious downside with this is extra copies in the interactive console path. So might it be worth exploring the posibility of a chardev multiplexor in QEMU. We would still pass in a pre-opened socketpair to QEMU for the logging side of things, but would leave the TCP/UNIX socket listening upto QEMU still. eg should we make something like this work: -add-fd set=2,fd=33 -chardev pipe,id=charserial0file,path=/dev/fdset/2 -chardev socket,id=charserial0tcp,host=127.0.0.1,port=9999,telnet,server,nowait -chardev multiplex,id=charserial0,muxA=charserial0file,muxB=charserial1 -serial isa-serial,chardev=charserial0,id=serial0 Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|

3 3

[libvirt] [PATCH v2 00/27] Prepare for specific vcpu hot(un)plug - part 1
by Peter Krempa 09 Dec '15

09 Dec '15

This series is getting rather big. The target is to refactor the way libvirt stores info about vCPUs into a single structure (okay, two structures for the qemu driver. Part 1 is not yet completely there, well, not even halfway. Future work will involve fully allocating priv->vcpupids to the maxcpus size and moving around few other bits of data in cputune and other parts to the new structure. Yet another follow up work is then to add new APIs for vCPU hotplug, which will enable adding vCPUs sparsely (useful if you have NUMA). Since this refactor will result in tracking all vcpu-related data in one struct, the result will automagically fix a few bugs where we'd end up with invalid config after vcpu unplug or other operations. Version 2 does not contain already pushed patches and incorporates feedback from John's review. Since I've changed quite a few things I'm reposting this. Peter Krempa (27): conf: Replace writes to def->maxvcpus with accessor conf: Use local copy of maxvcpus in virDomainVcpuParse conf: Extract update of vcpu count if maxvcpus is decreased conf: Add helper to check whether domain has offline vCPUs conf: Replace read access to def->maxvcpus with accessor conf: Replace writes to def->vcpus with accessor conf: Move vcpu count check into helper conf: Replace read accesses to def->vcpus with accessor conf: Turn def->maxvcpus into size_t qemu: domain: Add helper to access vm->privateData->agent qemu: Extract vCPU onlining/offlining via agent into a separate function qemu: qemuDomainSetVcpusAgent: re-check agent before calling it the again qemu: Split up vCPU hotplug and hotunplug qemu: cpu hotplug: Fix error handling logic qemu: monitor: Remove weird return values from qemuMonitorSetCPU qemu: cpu hotplug: Move loops to qemuDomainSetVcpusFlags qemu: Refactor qemuDomainHotplugVcpus qemu: refactor qemuDomainHotunplugVcpus conf: turn def->vcpus into a structure conf: ABI: Split up and improve vcpu info ABI checking conf: Add helper to get pointer to a certain vCPU definition qemu: cgroup: Remove now unreachable check qemu: Drop checking vcpu threads in emulator bandwidth getter/setter qemu: Replace checking for vcpu<->pid mapping availability with a helper qemu: Add helper to retrieve vCPU pid qemu: driver: Refactor qemuDomainHelperGetVcpus qemu: cgroup: Don't use priv->ncpupids to iterate domain vCPUs src/bhyve/bhyve_command.c | 2 +- src/bhyve/bhyve_driver.c | 2 +- src/conf/domain_audit.c | 2 +- src/conf/domain_conf.c | 190 ++++++++++++--- src/conf/domain_conf.h | 20 +- src/hyperv/hyperv_driver.c | 10 +- src/libvirt_private.syms | 6 + src/libxl/libxl_conf.c | 6 +- src/libxl/libxl_driver.c | 40 ++-- src/lxc/lxc_controller.c | 2 +- src/lxc/lxc_driver.c | 2 +- src/lxc/lxc_native.c | 7 +- src/openvz/openvz_conf.c | 7 +- src/openvz/openvz_driver.c | 19 +- src/phyp/phyp_driver.c | 12 +- src/qemu/qemu_cgroup.c | 42 ++-- src/qemu/qemu_command.c | 29 ++- src/qemu/qemu_domain.c | 47 ++++ src/qemu/qemu_domain.h | 4 + src/qemu/qemu_driver.c | 542 ++++++++++++++++++++++--------------------- src/qemu/qemu_monitor.c | 3 + src/qemu/qemu_monitor_json.c | 8 - src/qemu/qemu_monitor_text.c | 23 +- src/qemu/qemu_process.c | 22 +- src/test/test_driver.c | 38 +-- src/uml/uml_driver.c | 2 +- src/vbox/vbox_common.c | 19 +- src/vmware/vmware_driver.c | 2 +- src/vmx/vmx.c | 38 +-- src/vz/vz_driver.c | 8 +- src/vz/vz_sdk.c | 13 +- src/xen/xm_internal.c | 19 +- src/xenapi/xenapi_driver.c | 7 +- src/xenapi/xenapi_utils.c | 6 +- src/xenconfig/xen_common.c | 16 +- src/xenconfig/xen_sxpr.c | 27 ++- 36 files changed, 767 insertions(+), 475 deletions(-) -- 2.6.2

1 28

[libvirt] Handling large amount of VF's with intelligent passthrough
by Jan Gutter 09 Dec '15

09 Dec '15

Hi, I've run into a rather interesting problem recently where a weird interaction between libnl3 and libvirt caused some difficult-to-debug issues. From libvirt's side, the issue was that a netlink response was much larger than the pagesize and truncated by libnl3. When virNetDevLinkDump() calls virNetlinkCommand(), nl_recv() is supposed to return a rather large structure with information for all the Virtual Functions. When called in a system where the number of PCIe Virtual Functions are more than 30 for a given Physical Function, the netlink response is larger than 4k, meaning that a message is truncated. Unfortunately libnl3 truncates this silently, meaning that a cryptic error pops up much later in virNetDevParseVfConfig(), "missing IFLA_VF_INFO in netlink response". Aside from the error propagation (which might be fixable in libnl3), there still remains the need to enable libvirt to function in cases like this. This can be done in two ways, both in virNetlinkCommand(). 1. Message peeking can be enabled. In theory this slows down any netlink messages by doing a two stage query: query the buffer size, then allocate the receive buffer and receive the message. This is a reliability/performance tradeoff, I guess. This is as simple as adding: nl_socket_enable_msg_peek(nlhandle); 2. The receive buffer size can also be made larger: nl_socket_set_msg_buf_size(nlhandle, ARBITRARY_BUFFER_SIZE); This does not incur a performance penalty, but until libnl3 can propagate the truncation error, this merely postpones the error for future generations... Jan

3 5

[libvirt] [PATCH v3 0/6] Add read error ignore to storage backend vol info APIs
by John Ferlan 09 Dec '15

09 Dec '15

v2 here: http://www.redhat.com/archives/libvir-list/2015-November/thread.html Based on review and well just the amount of change I figured I'd generate a v3. The patches were mostly ACK'd, but patch 3 required removal of some flags which resulted in patch 1 needing updates... Then I found that what I'd done in patch 3 regarding returning errors needed to generate a patch 2.5... So I just bit the bullet and am posting the patches again. Patches 2, 4, and 5 of this series are unchanged. While patch 1 removes 'readflags' from an API, patch 3 is new, and patch 6 uses the new name. John Ferlan (6): storage: Add readflags for backend error processing storage: Add comments for backend APIs storage: Set ret = -1 on failures in virStorageBackendUpdateVolTargetInfo storage: Handle readflags errors storage: Add debug message storage: Ignore block devices that fail format detection src/storage/storage_backend.c | 83 ++++++++++++++++++++++++++++++----- src/storage/storage_backend.h | 15 ++++++- src/storage/storage_backend_disk.c | 5 ++- src/storage/storage_backend_fs.c | 4 +- src/storage/storage_backend_logical.c | 2 +- src/storage/storage_backend_mpath.c | 2 +- src/storage/storage_backend_scsi.c | 6 ++- 7 files changed, 96 insertions(+), 21 deletions(-) -- 2.5.0

2 7

[libvirt] [PATCH v2] guest-agent-socket: don't generate default path to config XML
by Pavel Hrdina 09 Dec '15

09 Dec '15

While we started using for all unix sockets as default one common directory based on a guest name it introduced several issues, for example with renaming the guest or cloning it. In general it's not entirely bad, but in this case it would be best to hide the auto-generated socket path from user and don't export it in the config XML. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- Version 2: - removed unnecessary changes src/qemu/qemu_command.c | 28 ++++++++++++++++++++++++++++ src/qemu/qemu_command.h | 1 + src/qemu/qemu_domain.c | 16 ---------------- src/qemu/qemu_driver.c | 3 +++ src/qemu/qemu_process.c | 3 +++ tests/qemuhotplugtest.c | 5 +++++ tests/qemuxml2argvtest.c | 5 +++++ 7 files changed, 45 insertions(+), 16 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 2a9fab5..f1bb621 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -1262,6 +1262,34 @@ qemuAssignDeviceAliases(virDomainDefPtr def, virQEMUCapsPtr qemuCaps) } +int +qemuUnixSocketGenerate(virDomainDefPtr def, + virQEMUDriverConfigPtr cfg) +{ + size_t i; + + for (i = 0; i < def->nchannels; i++) { + virDomainChrDefPtr channel = def->channels[i]; + + if (channel->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_CHANNEL && + channel->targetType == VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO && + channel->source.type == VIR_DOMAIN_CHR_TYPE_UNIX && + !channel->source.data.nix.path) { + if (virAsprintf(&channel->source.data.nix.path, + "%s/domain-%s/%s", + cfg->channelTargetDir, def->name, + channel->target.name ? channel->target.name + : "unknown.sock") < 0) + return -1; + + channel->source.data.nix.listen = true; + } + } + + return 0; +} + + static void qemuDomainPrimeVirtioDeviceAddresses(virDomainDefPtr def, virDomainDeviceAddressType type) diff --git a/src/qemu/qemu_command.h b/src/qemu/qemu_command.h index bebdd27..0512a23 100644 --- a/src/qemu/qemu_command.h +++ b/src/qemu/qemu_command.h @@ -283,6 +283,7 @@ int qemuAssignDevicePCISlots(virDomainDefPtr def, virDomainPCIAddressSetPtr addrs); int qemuAssignDeviceAliases(virDomainDefPtr def, virQEMUCapsPtr qemuCaps); +int qemuUnixSocketGenerate(virDomainDefPtr def, virQEMUDriverConfigPtr cfg); int qemuDomainNetVLAN(virDomainNetDefPtr def); int qemuAssignDeviceNetAlias(virDomainDefPtr def, virDomainNetDefPtr net, int idx); int qemuAssignDeviceDiskAlias(virDomainDefPtr vmdef, diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index ed21245..7e05289 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -1329,22 +1329,6 @@ qemuDomainDeviceDefPostParse(virDomainDeviceDefPtr dev, ARCH_IS_S390(def->os.arch)) dev->data.controller->model = VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI; - /* auto generate unix socket path */ - if (dev->type == VIR_DOMAIN_DEVICE_CHR && - dev->data.chr->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_CHANNEL && - dev->data.chr->targetType == VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO && - dev->data.chr->source.type == VIR_DOMAIN_CHR_TYPE_UNIX && - !dev->data.chr->source.data.nix.path) { - if (virAsprintf(&dev->data.chr->source.data.nix.path, - "%s/domain-%s/%s", - cfg->channelTargetDir, def->name, - dev->data.chr->target.name ? dev->data.chr->target.name - : "unknown.sock") < 0) - goto cleanup; - - dev->data.chr->source.data.nix.listen = true; - } - /* forbid capabilities mode hostdev in this kind of hypervisor */ if (dev->type == VIR_DOMAIN_DEVICE_HOSTDEV && dev->data.hostdev->mode == VIR_DOMAIN_HOSTDEV_MODE_CAPABILITIES) { diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index ae1d8e7..39f2f03 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -7254,6 +7254,9 @@ static char *qemuConnectDomainXMLToNative(virConnectPtr conn, if (qemuAssignDeviceAliases(def, qemuCaps) < 0) goto cleanup; + if (qemuUnixSocketGenerate(def, cfg) < 0) + goto cleanup; + if (qemuDomainAssignAddresses(def, qemuCaps, NULL) < 0) goto cleanup; diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 192730c..29cf965 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -4663,6 +4663,9 @@ qemuProcessLaunch(virConnectPtr conn, if (qemuAssignDeviceAliases(vm->def, priv->qemuCaps) < 0) goto cleanup; + if (qemuUnixSocketGenerate(vm->def, cfg) < 0) + goto cleanup; + /* Get the advisory nodeset from numad if 'placement' of * either <vcpu> or <numatune> is 'auto'. */ diff --git a/tests/qemuhotplugtest.c b/tests/qemuhotplugtest.c index 102e052..b17cca2 100644 --- a/tests/qemuhotplugtest.c +++ b/tests/qemuhotplugtest.c @@ -61,6 +61,7 @@ qemuHotplugCreateObjects(virDomainXMLOptionPtr xmlopt, { int ret = -1; qemuDomainObjPrivatePtr priv = NULL; + virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(&driver); if (!(*vm = virDomainObjNew(xmlopt))) goto cleanup; @@ -94,10 +95,14 @@ qemuHotplugCreateObjects(virDomainXMLOptionPtr xmlopt, if (qemuAssignDeviceAliases((*vm)->def, priv->qemuCaps) < 0) goto cleanup; + if (qemuUnixSocketGenerate((*vm)->def, cfg) < 0) + goto cleanup; + (*vm)->def->id = QEMU_HOTPLUG_TEST_DOMAIN_ID; ret = 0; cleanup: + virObjectUnref(cfg); return ret; } diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index f7596a0..f6083db 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -262,6 +262,7 @@ static int testCompareXMLToArgvFiles(const char *xml, virCommandPtr cmd = NULL; size_t i; virBitmapPtr nodeset = NULL; + virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(&driver); if (!(conn = virGetConnect())) goto out; @@ -324,6 +325,9 @@ static int testCompareXMLToArgvFiles(const char *xml, if (qemuAssignDeviceAliases(vmdef, extraFlags) < 0) goto out; + if (qemuUnixSocketGenerate(vmdef, cfg) < 0) + goto out; + for (i = 0; i < vmdef->nhostdevs; i++) { virDomainHostdevDefPtr hostdev = vmdef->hostdevs[i]; @@ -387,6 +391,7 @@ static int testCompareXMLToArgvFiles(const char *xml, virCommandFree(cmd); virDomainDefFree(vmdef); virObjectUnref(conn); + virObjectUnref(cfg); virBitmapFree(nodeset); return ret; } -- 2.6.3

2 1

[libvirt] somtimes libvirtd crashed when fail to start a guest, need help to debug
by Luyao Huang 09 Dec '15

09 Dec '15

Dear list, I met a problem when test with the latest libvirt (build from current master branch), i found sometimes libvirtd crashed when guest fail to start. the reproduce way is very easy, but hard to meet: # cat /root/shdo2.sh #！/bin/sh while : do virsh start rhel7.0-rhel done # sh /root/shdo2.sh ... error: Failed to start domain rhel7.0-rhel error: internal error: process exited while connecting to monitor: 2015-12-09T06:58:37.785695Z qemu-kvm: -chardev socket,id=charchannel1,path=/var/lib/libvirt/qemu1/rhel7.0-rhel.agent,server,nowait: Failed to bind socket to /var/lib/libvirt/qemu1/rhel7.0-rhel.agent: No such file or directory error: Failed to start domain rhel7.0-rhel error: internal error: process exited while connecting to monitor: 2015-12-09T06:58:42.035103Z qemu-kvm: -chardev socket,id=charchannel1,path=/var/lib/libvirt/qemu1/rhel7.0-rhel.agent,server,nowait: Failed to bind socket to /var/lib/libvirt/qemu1/rhel7.0-rhel.agent: No such file or directory error: Disconnected from qemu:///system due to I/O error error: Failed to start domain rhel7.0-rhel error: End of file while reading data: Input/output error error: One or more references were leaked after disconnect from the hypervisor error: failed to connect to the hypervisor error: no valid connection error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': Connection refused and here is the bt full output: http://pastebin.com/sCYHByvP and here is the valgrind output: http://pastebin.com/0VXrYTKT I cannot found what happened, also i am not familiar with virtlogd. could someone please help to debug this ? Thanks, Luyao

1 0

[libvirt] [PATCH v2] log: include hostname in initial log message
by Daniel P. Berrange 09 Dec '15

09 Dec '15

On the very first log message we send to any output, we include the libvirt version number and package string. In some bug reports we have been given libvirtd.log files that came from a different host than the corresponding /var/log/libvirt/qemu log files. So extend the initial log message to include the hostname too. eg on first log message we would now see: $ libvirtd 2015-12-04 17:35:36.610+0000: 20917: info : libvirt version: 1.3.0 2015-12-04 17:35:36.610+0000: 20917: info : hostname: dhcp-1-180.lcy.redhat.com 2015-12-04 17:35:36.610+0000: 20917: error : qemuMonitorIO:687 : internal error: End of file from monitor Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/libvirt_private.syms | 1 + src/util/virlog.c | 72 ++++++++++++++++++++++++++++++++++++------------ src/util/virutil.c | 34 ++++++++++++++++------- src/util/virutil.h | 1 + 4 files changed, 81 insertions(+), 27 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index dd085c3..f483c4c 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -2387,6 +2387,7 @@ virGetGroupID; virGetGroupList; virGetGroupName; virGetHostname; +virGetHostnameQuiet; virGetListenFDs; virGetSCSIHostNameByParentaddr; virGetSCSIHostNumber; diff --git a/src/util/virlog.c b/src/util/virlog.c index 627f4cb..7d5a266 100644 --- a/src/util/virlog.c +++ b/src/util/virlog.c @@ -94,7 +94,7 @@ static int virLogNbFilters; * after filtering, multiple output can be used simultaneously */ struct _virLogOutput { - bool logVersion; + bool logInitMessage; void *data; virLogOutputFunc f; virLogCloseFunc c; @@ -402,7 +402,7 @@ virLogDefineOutput(virLogOutputFunc f, goto cleanup; } ret = virLogNbOutputs++; - virLogOutputs[ret].logVersion = true; + virLogOutputs[ret].logInitMessage = true; virLogOutputs[ret].f = f; virLogOutputs[ret].c = c; virLogOutputs[ret].data = data; @@ -452,6 +452,32 @@ virLogVersionString(const char **rawmsg, return virLogFormatString(msg, 0, NULL, VIR_LOG_INFO, VIR_LOG_VERSION_STRING); } +/* Similar to virGetHostname() but avoids use of error + * reporting APIs or logging APIs, to prevent recursion + */ +static int +virLogHostnameString(const char **rawmsg, + char **msg) +{ + char *hostname = virGetHostnameQuiet(); + char *hoststr; + + if (!hostname) + return -1; + + if (virAsprintfQuiet(&hoststr, "hostname: %s", hostname) < 0) { + VIR_FREE(hostname); + return -1; + } + + if (virLogFormatString(msg, 0, NULL, VIR_LOG_INFO, hoststr) < 0) { + VIR_FREE(hoststr); + return -1; + } + *rawmsg = hoststr; + return 0; +} + static void virLogSourceUpdate(virLogSourcePtr source) @@ -533,7 +559,7 @@ virLogVMessage(virLogSourcePtr source, const char *fmt, va_list vargs) { - static bool logVersionStderr = true; + static bool logInitMessageStderr = true; char *str = NULL; char *msg = NULL; char timestamp[VIR_TIME_STRING_BUFLEN]; @@ -583,16 +609,22 @@ virLogVMessage(virLogSourcePtr source, */ for (i = 0; i < virLogNbOutputs; i++) { if (priority >= virLogOutputs[i].priority) { - if (virLogOutputs[i].logVersion) { - const char *rawver; - char *ver = NULL; - if (virLogVersionString(&rawver, &ver) >= 0) + if (virLogOutputs[i].logInitMessage) { + const char *rawinitmsg; + char *initmsg = NULL; + if (virLogVersionString(&rawinitmsg, &initmsg) >= 0) + virLogOutputs[i].f(&virLogSelf, VIR_LOG_INFO, + __FILE__, __LINE__, __func__, + timestamp, NULL, 0, rawinitmsg, initmsg, + virLogOutputs[i].data); + VIR_FREE(initmsg); + if (virLogHostnameString(&rawinitmsg, &initmsg) >= 0) virLogOutputs[i].f(&virLogSelf, VIR_LOG_INFO, __FILE__, __LINE__, __func__, - timestamp, NULL, 0, rawver, ver, + timestamp, NULL, 0, rawinitmsg, initmsg, virLogOutputs[i].data); - VIR_FREE(ver); - virLogOutputs[i].logVersion = false; + VIR_FREE(initmsg); + virLogOutputs[i].logInitMessage = false; } virLogOutputs[i].f(source, priority, filename, linenr, funcname, @@ -601,16 +633,22 @@ virLogVMessage(virLogSourcePtr source, } } if (virLogNbOutputs == 0) { - if (logVersionStderr) { - const char *rawver; - char *ver = NULL; - if (virLogVersionString(&rawver, &ver) >= 0) + if (logInitMessageStderr) { + const char *rawinitmsg; + char *initmsg = NULL; + if (virLogVersionString(&rawinitmsg, &initmsg) >= 0) + virLogOutputToFd(&virLogSelf, VIR_LOG_INFO, + __FILE__, __LINE__, __func__, + timestamp, NULL, 0, rawinitmsg, initmsg, + (void *) STDERR_FILENO); + VIR_FREE(initmsg); + if (virLogHostnameString(&rawinitmsg, &initmsg) >= 0) virLogOutputToFd(&virLogSelf, VIR_LOG_INFO, __FILE__, __LINE__, __func__, - timestamp, NULL, 0, rawver, ver, + timestamp, NULL, 0, rawinitmsg, initmsg, (void *) STDERR_FILENO); - VIR_FREE(ver); - logVersionStderr = false; + VIR_FREE(initmsg); + logInitMessageStderr = false; } virLogOutputToFd(source, priority, filename, linenr, funcname, diff --git a/src/util/virutil.c b/src/util/virutil.c index fe65faf..e83def1 100644 --- a/src/util/virutil.c +++ b/src/util/virutil.c @@ -662,16 +662,17 @@ char *virIndexToDiskName(int idx, const char *prefix) * we got from getaddrinfo(). Return the value from gethostname() * and hope for the best. */ -char *virGetHostname(void) +static char *virGetHostnameImpl(bool quiet) { int r; - char hostname[HOST_NAME_MAX+1], *result; + char hostname[HOST_NAME_MAX+1], *result = NULL; struct addrinfo hints, *info; r = gethostname(hostname, sizeof(hostname)); if (r == -1) { - virReportSystemError(errno, - "%s", _("failed to determine host name")); + if (!quiet) + virReportSystemError(errno, + "%s", _("failed to determine host name")); return NULL; } NUL_TERMINATE(hostname); @@ -683,7 +684,7 @@ char *virGetHostname(void) * string as-is; it's up to callers to check whether "localhost" * is allowed. */ - ignore_value(VIR_STRDUP(result, hostname)); + ignore_value(VIR_STRDUP_QUIET(result, hostname)); goto cleanup; } @@ -696,9 +697,10 @@ char *virGetHostname(void) hints.ai_family = AF_UNSPEC; r = getaddrinfo(hostname, NULL, &hints, &info); if (r != 0) { - VIR_WARN("getaddrinfo failed for '%s': %s", - hostname, gai_strerror(r)); - ignore_value(VIR_STRDUP(result, hostname)); + if (!quiet) + VIR_WARN("getaddrinfo failed for '%s': %s", + hostname, gai_strerror(r)); + ignore_value(VIR_STRDUP_QUIET(result, hostname)); goto cleanup; } @@ -711,17 +713,29 @@ char *virGetHostname(void) * localhost. Ignore the canonicalized name and just return the * original hostname */ - ignore_value(VIR_STRDUP(result, hostname)); + ignore_value(VIR_STRDUP_QUIET(result, hostname)); else /* Caller frees this string. */ - ignore_value(VIR_STRDUP(result, info->ai_canonname)); + ignore_value(VIR_STRDUP_QUIET(result, info->ai_canonname)); freeaddrinfo(info); cleanup: + if (!result) { + virReportOOMError(); + } return result; } +char *virGetHostname(void) +{ + return virGetHostnameImpl(false); +} + +char *virGetHostnameQuiet(void) +{ + return virGetHostnameImpl(true); +} char * virGetUserDirectory(void) diff --git a/src/util/virutil.h b/src/util/virutil.h index 02387e0..535807c 100644 --- a/src/util/virutil.h +++ b/src/util/virutil.h @@ -132,6 +132,7 @@ static inline int pthread_sigmask(int how, # endif char *virGetHostname(void); +char *virGetHostnameQuiet(void); char *virGetUserDirectory(void); char *virGetUserDirectoryByUID(uid_t uid); -- 2.5.0

2 1

[libvirt] Event after reverting to snapshot
by Dmitry Andreev 09 Dec '15

09 Dec '15

Hi, I'm looking for a way to catch an event about reverting to snapshot. As I can see in the code there is no lifecycle event if VM state wasn't changed. The comment in the code of qemuDomainRevertToSnapshot: /* We have the following transitions, which create the following events: * 1. inactive -> inactive: none * 2. inactive -> running: EVENT_STARTED * 3. inactive -> paused: EVENT_STARTED, EVENT_PAUSED * 4. running -> inactive: EVENT_STOPPED * 5. running -> running: none * 6. running -> paused: EVENT_PAUSED * 7. paused -> inactive: EVENT_STOPPED * 8. paused -> running: EVENT_RESUMED * 9. paused -> paused: none * Also, several transitions occur even if we fail partway through, * and use of FORCE can cause multiple transitions. */ Also there is no VIR_DOMAIN_EVENT_DEFINED event after reverting to snapshot, even when domain configuration is changed. Isn't that strange? Checked on stopped VM. Is there any way to catch event about changes in configuration after reverting to snapshot.

2 1