- Devel - libvirt List Archives

[libvirt] [PATCH] network: Defer online of macvtap during qemu migration
by Matthew Rosato 14 May '14

14 May '14

When generating macvtaps via virNetDevMacVLanCreateWithVPortProfile, the macvtap device is unconditionally set to the up state. However, during migration, this results in a case where both the source and target system are simultaneously up with the same MAC address. This patch defers bringing the target macvtap up until later in the migration to shrink this window. Signed-off-by: Matthew Rosato <mjrosato(a)linux.vnet.ibm.com> --- src/qemu/qemu_migration.c | 18 ++++++++++++++++++ src/util/virnetdevmacvlan.c | 11 ++++++++--- 2 files changed, 26 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index a9f7fea..aee803a 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -56,6 +56,7 @@ #include "virhook.h" #include "virstring.h" #include "virtypedparam.h" +#include "virnetdev.h" #define VIR_FROM_THIS VIR_FROM_QEMU @@ -4468,6 +4469,8 @@ qemuMigrationFinish(virQEMUDriverPtr driver, virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver); virCapsPtr caps = NULL; unsigned short port; + virDomainNetDefPtr net; + size_t i; VIR_DEBUG("driver=%p, dconn=%p, vm=%p, cookiein=%s, cookieinlen=%d, " "cookieout=%p, cookieoutlen=%p, flags=%lx, retcode=%d", @@ -4574,6 +4577,21 @@ qemuMigrationFinish(virQEMUDriverPtr driver, } if (!(flags & VIR_MIGRATE_PAUSED) && !(flags & VIR_MIGRATE_OFFLINE)) { + /* Macvtaps were previously left offline, bring them online now */ + for (i = 0; i < vm->def->nnets; i++) { + net = vm->def->nets[i]; + if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) { + if (virNetDevSetOnline(net->ifname, true) < 0) { + ignore_value(virNetDevVPortProfileDisassociate(net->ifname, + virDomainNetGetActualVirtPortProfile(net), + &net->mac, + virDomainNetGetActualDirectDev(net), + -1, + VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH)); + ignore_value(virNetDevMacVLanDelete(net->ifname)); + } + } + } /* run 'cont' on the destination, which allows migration on qemu * >= 0.10.6 to work properly. This isn't strictly necessary on * older qemu's, but it also doesn't hurt anything there diff --git a/src/util/virnetdevmacvlan.c b/src/util/virnetdevmacvlan.c index 7bbf540..3da845b 100644 --- a/src/util/virnetdevmacvlan.c +++ b/src/util/virnetdevmacvlan.c @@ -898,9 +898,14 @@ int virNetDevMacVLanCreateWithVPortProfile(const char *tgifname, goto link_del_exit; } - if (virNetDevSetOnline(cr_ifname, true) < 0) { - rc = -1; - goto disassociate_exit; + /* If this device is being created as part of an inbound + * migration, leave the device offline for now. + */ + if (vmOp != VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_START) { + if (virNetDevSetOnline(cr_ifname, true) < 0) { + rc = -1; + goto disassociate_exit; + } } if (withTap) { -- 1.7.9.5

3 5

[libvirt] [PATCH] qemu: snapshot: Terminate job when memory compression program isn't found
by Peter Krempa 14 May '14

14 May '14

If the compression program for external snapshot memory image isn't found we exitted the function without terminating the domain job. This caused the domain to be unusable. The problem was introduced in commit 7df5093f. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1097503 --- src/qemu/qemu_driver.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index a5c7f4f..52ca47c 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -13207,13 +13207,14 @@ qemuDomainSnapshotCreateActiveExternal(virConnectPtr conn, virReportError(VIR_ERR_OPERATION_FAILED, "%s", _("Invalid snapshot image format specified " "in configuration file")); - goto cleanup; + goto endjob; } + if (!qemuCompressProgramAvailable(compressed)) { virReportError(VIR_ERR_OPERATION_FAILED, "%s", _("Compression program for image format " "in configuration file isn't available")); - goto cleanup; + goto endjob; } } -- 1.9.2

2 2

[libvirt] daemon-config-network dependency
by Cedric Bosdonnat 14 May '14

14 May '14

Hi all, The libvirt-daemon-config-network package is only pulled by libvirt. So a user installing only libvirt-daemon-lxc (qemu, xen, etc) won't get that package... and no default network. Would it be OK if I move the dependency to libvirt-daemon-{lxc,qemu,xen,...} packages? -- Cedric

2 2

Re: [libvirt] [Qemu-devel] qemu leaving unix sockets behind after VM is shut down
by Stefan Hajnoczi 13 May '14

13 May '14

On Tue, Apr 01, 2014 at 02:34:58PM -0600, Chris Friesen wrote: > When running qemu with something like this > > -device virtio-serial \ > -chardev socket,path=/tmp/foo,server,nowait,id=foo \ > -device virtserialport,chardev=foo,name=host.port.0 > > the VM starts up as expected and creates a socket at /tmp/foo as expected. > > However, when I shut down the VM the socket at /tmp/foo is left > behind in the filesystem. Basically qemu has "leaked" a file. > > With something like OpenStack where we could be creating/destroying > many VMs this could end up creating a significant number of files in > the specified directory. > > Has any thought been given to either automatically cleaning up the > unix socket in the filesystem when qemu exits, or else supporting > the abstract namespace for unix sockets to allow for automatic > cleanup? Libvirt has a special case for the monitor socket in its qemuProcessStop() function. Are you using the OpenStack libvirt driver? Perhaps QEMU should support cleanup but first I think we should check the situation with libvirt. Stefan

5 4

[libvirt] [libvirt-python PATCH] override: add virDomainFSFreeze and virDomainFSThaw API
by Tomoki Sekiyama 13 May '14

13 May '14

Add binding for the new virDomainFSFreeze and virDomainFSThaw functions added in libvirt 1.2.5. These require override since these take a list of mountpoints path string. The methods are named 'fSFreeze' and 'fSThaw', like a existing 'fSTrim' method. Signed-off-by: Tomoki Sekiyama <tomoki.sekiyama(a)hds.com> --- generator.py | 2 + libvirt-override-api.xml | 14 +++++++ libvirt-override.c | 97 ++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 113 insertions(+) diff --git a/generator.py b/generator.py index 05ec743..eec81b0 100755 --- a/generator.py +++ b/generator.py @@ -462,6 +462,8 @@ skip_impl = ( 'virDomainMigrate3', 'virDomainMigrateToURI3', 'virConnectGetCPUModelNames', + 'virDomainFSFreeze', + 'virDomainFSThaw', ) lxc_skip_impl = ( diff --git a/libvirt-override-api.xml b/libvirt-override-api.xml index d5b25b5..a3db33e 100644 --- a/libvirt-override-api.xml +++ b/libvirt-override-api.xml @@ -624,5 +624,19 @@ <arg name='conn' type='virConnectPtr' info='pointer to the hypervisor connection'/> <arg name='flags' type='int' info='unused, pass 0'/> </function> + <function name='virDomainFSFreeze' file='python'> + <info>Freeze specified filesystems within the guest</info> + <return type='int' info='the number of frozen filesystems on success, -1 otherwise.'/> + <arg name='dom' type='virDomainPtr' info='a domain object'/> + <arg name='mountpoints' type='const char **' info='list of mount points to be frozen, or None'/> + <arg name='flags' type='unsigned int' info='unused, pass 0'/> + </function> + <function name='virDomainFSThaw' file='python'> + <info>Thaw specified filesystems within the guest</info> + <return type='int' info='the number of thawed filesystems on success, -1 otherwise.'/> + <arg name='dom' type='virDomainPtr' info='a domain object'/> + <arg name='mountpoints' type='const char **' info='list of mount points to be thawed, or None'/> + <arg name='flags' type='unsigned int' info='unused, pass 0'/> + </function> </symbols> </api> diff --git a/libvirt-override.c b/libvirt-override.c index 3fa9b9b..d0557c2 100644 --- a/libvirt-override.c +++ b/libvirt-override.c @@ -7554,6 +7554,99 @@ cleanup: #endif /* LIBVIR_CHECK_VERSION(1, 1, 1) */ +#if LIBVIR_CHECK_VERSION(1, 2, 5) +static PyObject * +libvirt_virDomainFSFreeze(PyObject *self ATTRIBUTE_UNUSED, PyObject *args) { + PyObject *py_retval = NULL; + int c_retval; + virDomainPtr domain; + PyObject *pyobj_domain; + PyObject *pyobj_list; + unsigned int flags; + unsigned int nmountpoints = 0; + const char **mountpoints = NULL; + size_t i = 0, j; + + if (!PyArg_ParseTuple(args, (char *)"OOi:virDomainFSFreeze", + &pyobj_domain, &pyobj_list, &flags)) + return NULL; + domain = (virDomainPtr) PyvirDomain_Get(pyobj_domain); + + if (PyList_Check(pyobj_list)) { + nmountpoints = PyList_Size(pyobj_list); + + if (VIR_ALLOC_N(mountpoints, nmountpoints) < 0) + return PyErr_NoMemory(); + + for (i = 0; i < nmountpoints; i++) { + if (libvirt_charPtrUnwrap(PyList_GetItem(pyobj_list, i), + (char **)mountpoints+i) < 0 || + mountpoints[i] == NULL) + goto cleanup; + } + } + + LIBVIRT_BEGIN_ALLOW_THREADS; + c_retval = virDomainFSFreeze(domain, mountpoints, nmountpoints, flags); + LIBVIRT_END_ALLOW_THREADS; + py_retval = libvirt_intWrap((int) c_retval); + +cleanup: + if (mountpoints) { + for (j = 0 ; j < i ; j++) + VIR_FREE(mountpoints[j]); + VIR_FREE(mountpoints); + } + return py_retval; +} + + +static PyObject * +libvirt_virDomainFSThaw(PyObject *self ATTRIBUTE_UNUSED, PyObject *args) { + PyObject *py_retval = NULL; + int c_retval; + virDomainPtr domain; + PyObject *pyobj_domain; + PyObject *pyobj_list; + unsigned int flags; + unsigned int nmountpoints = 0; + const char **mountpoints = NULL; + size_t i = 0, j; + + if (!PyArg_ParseTuple(args, (char *)"OOi:virDomainFSThaw", + &pyobj_domain, &pyobj_list, &flags)) + return NULL; + domain = (virDomainPtr) PyvirDomain_Get(pyobj_domain); + + if (PyList_Check(pyobj_list)) { + nmountpoints = PyList_Size(pyobj_list); + + if (VIR_ALLOC_N(mountpoints, nmountpoints) < 0) + return PyErr_NoMemory(); + + for (i = 0; i < nmountpoints; i++) { + if (libvirt_charPtrUnwrap(PyList_GetItem(pyobj_list, i), + (char **)mountpoints+i) < 0 || + mountpoints[i] == NULL) + goto cleanup; + } + } + + LIBVIRT_BEGIN_ALLOW_THREADS; + c_retval = virDomainFSThaw(domain, mountpoints, nmountpoints, flags); + LIBVIRT_END_ALLOW_THREADS; + py_retval = libvirt_intWrap((int) c_retval); + +cleanup: + if (mountpoints) { + for (j = 0 ; j < i ; j++) + VIR_FREE(mountpoints[j]); + VIR_FREE(mountpoints); + } + return py_retval; +} +#endif /* LIBVIR_CHECK_VERSION(1, 2, 5) */ + /************************************************************************ * * * The registration stuff * @@ -7729,6 +7822,10 @@ static PyMethodDef libvirtMethods[] = { {(char *) "virDomainCreateXMLWithFiles", libvirt_virDomainCreateXMLWithFiles, METH_VARARGS, NULL}, {(char *) "virDomainCreateWithFiles", libvirt_virDomainCreateWithFiles, METH_VARARGS, NULL}, #endif /* LIBVIR_CHECK_VERSION(1, 1, 1) */ +#if LIBVIR_CHECK_VERSION(1, 2, 5) + {(char *) "virDomainFSFreeze", libvirt_virDomainFSFreeze, METH_VARARGS, NULL}, + {(char *) "virDomainFSThaw", libvirt_virDomainFSThaw, METH_VARARGS, NULL}, +#endif /* LIBVIR_CHECK_VERSION(1, 2, 5) */ {NULL, NULL, 0, NULL} };

4 5

[libvirt] [Question] (python API) setMemoryParameters doesn't support flags CONFIG | LIVE
by Wangrui (K) 13 May '14

13 May '14

Hi, all when I use (python API) setMemoryParameters with flags= VIR_DOMAIN_AFFECT_CONFIG|VIR_DOMAIN_AFFECT_LIVE , an error occurs as follow: (libvirt 1.2.4) libvirt: error : flags 'affect live' and 'affect config' in virDomainGetMemoryParameters are mutually exclusive I check the function libvirt_virDomainSetMemoryParameters in libvirt-override.c and find the code below: Line 1204-1206: LIBVIRT_BEGIN_ALLOW_THREADS; i_retval = virDomainGetMemoryParameters(domain, NULL, &nparams, flags); LIBVIRT_END_ALLOW_THREADS; Line 1220-1222: LIBVIRT_BEGIN_ALLOW_THREADS; i_retval = virDomainGetMemoryParameters(domain, params, &nparams, flags); LIBVIRT_END_ALLOW_THREADS; Function virDomainGetMemoryParameters doesn't support flags VIR_DOMAIN_AFFECT_CONFIG|VIR_DOMAIN_AFFECT_LIVE. So the error above occurs. The first call of virDomainGetMemoryParameters is to "cause @nparams on output to contain the number of parameters supported by the hypervisor". And what's the role of the second call? Maybe, so do other functions such as *libvirt_virDomainSetSchedulerParameters, *libvirt_virDomainSetBlkioParameters In commit 56cec18d761a1f99862c43811de60380c65881e6. IMHO, an example flag (such as VIR_DOMAIN_AFFECT_CONFIG) can be used in virDomainGetMemoryParameters just to get something from hypervisor rather than @flags which is a parameter of setMemoryParameters .

1 0

[libvirt] [PATCH] Improve the error info when fail to start guest due to lack of cgroup controller
by Shanzhi Yu 13 May '14

13 May '14

When cgroup controller cpu, cpuset and cpuacct are all unmounted, fail to start guest. The error info is not accurate. It will report "At least one cgroup controller is required" even though there is cgroup controller memory or blkio. Signed-off-by: Shanzhi Yu <shyu(a)redhat.com> --- src/util/vircgroup.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index fce380a..c609eb3 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -586,6 +586,11 @@ virCgroupDetect(virCgroupPtr group, if (!group->controllers[i].mountPoint) { VIR_DEBUG("Requested controller '%s' not mounted, ignoring", virCgroupControllerTypeToString(i)); + if ( i==2 ) { + virReportSystemError(ENXIO, "%s", + _("At least one cgroup controller cpu or cpuacct or cpuset is required")); + return -1; + } controllers &= ~(1 << i); } } else { -- 1.8.3.1

1 0

[libvirt] [PATCH] sanlock: avoid leak in acquire()
by Martin Kletzander 13 May '14

13 May '14

Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- src/locking/lock_driver_sanlock.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/locking/lock_driver_sanlock.c b/src/locking/lock_driver_sanlock.c index b3d31bf..e4b4e8e 100644 --- a/src/locking/lock_driver_sanlock.c +++ b/src/locking/lock_driver_sanlock.c @@ -907,9 +907,6 @@ static int virLockManagerSanlockAcquire(virLockManagerPtr lock, return -1; } - if (VIR_ALLOC(opt) < 0) - return -1; - /* We only initialize 'sock' if we are in the real * child process and we need it to be inherited * @@ -944,6 +941,9 @@ static int virLockManagerSanlockAcquire(virLockManagerPtr lock, return 0; } + if (VIR_ALLOC(opt) < 0) + return -1; + /* sanlock doesn't use owner_name for anything, so it's safe to take just * the first SANLK_NAME_LEN - 1 characters from vm_name */ ignore_value(virStrncpy(opt->owner_name, priv->vm_name, -- 1.9.3

2 2

[libvirt] [PATCH 1/2] sanlock: code movement in virLockManagerSanlockAcquire
by Martin Kletzander 13 May '14

13 May '14

Just move some code around for future patches to ease the review. With this patch there is no need for drastic cleanup path later. Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- src/locking/lock_driver_sanlock.c | 50 +++++++++++++++++++-------------------- 1 file changed, 25 insertions(+), 25 deletions(-) diff --git a/src/locking/lock_driver_sanlock.c b/src/locking/lock_driver_sanlock.c index 7e4258d..01441a0 100644 --- a/src/locking/lock_driver_sanlock.c +++ b/src/locking/lock_driver_sanlock.c @@ -896,31 +896,6 @@ static int virLockManagerSanlockAcquire(virLockManagerPtr lock, if (VIR_ALLOC(opt) < 0) return -1; - /* sanlock doesn't use owner_name for anything, so it's safe to take just - * the first SANLK_NAME_LEN - 1 characters from vm_name */ - ignore_value(virStrncpy(opt->owner_name, priv->vm_name, - SANLK_NAME_LEN - 1, SANLK_NAME_LEN)); - - if (state && STRNEQ(state, "")) { - if ((rv = sanlock_state_to_args((char *)state, - &res_count, - &res_args)) < 0) { - if (rv <= -200) - virReportError(VIR_ERR_INTERNAL_ERROR, - _("Unable to parse lock state %s: error %d"), - state, rv); - else - virReportSystemError(-rv, - _("Unable to parse lock state %s"), - state); - goto error; - } - res_free = true; - } else { - res_args = priv->res_args; - res_count = priv->res_count; - } - /* We only initialize 'sock' if we are in the real * child process and we need it to be inherited * @@ -949,6 +924,31 @@ static int virLockManagerSanlockAcquire(virLockManagerPtr lock, } } + /* sanlock doesn't use owner_name for anything, so it's safe to take just + * the first SANLK_NAME_LEN - 1 characters from vm_name */ + ignore_value(virStrncpy(opt->owner_name, priv->vm_name, + SANLK_NAME_LEN - 1, SANLK_NAME_LEN)); + + if (state && STRNEQ(state, "")) { + if ((rv = sanlock_state_to_args((char *)state, + &res_count, + &res_args)) < 0) { + if (rv <= -200) + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Unable to parse lock state %s: error %d"), + state, rv); + else + virReportSystemError(-rv, + _("Unable to parse lock state %s"), + state); + goto error; + } + res_free = true; + } else { + res_args = priv->res_args; + res_count = priv->res_count; + } + if (!(flags & VIR_LOCK_MANAGER_ACQUIRE_REGISTER_ONLY)) { VIR_DEBUG("Acquiring object %u", priv->res_count); if ((rv = sanlock_acquire(sock, priv->vm_pid, 0, -- 1.9.2

3 5

[libvirt] [PATCH] apibuild: Disallow 'returns' return decription
by Michal Privoznik 13 May '14

13 May '14

Our documentation generator is a bit messy, to say the least. For instance, the description to return values of a function is searched within C comment. Currently, all lines that start with 'returns' or 'Returns' are viewed as return value description. However, there are some valid uses where the 'returns' word is in the middle of a sentence describing function behavior nor the return value. For instance: virDomainDetachDeviceFlags, virConnectNetworkEventRegisterAny and virDomainGetDiskErrors. This leads to HTML documentation to be not correctly generated. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- docs/apibuild.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/apibuild.py b/docs/apibuild.py index 71494d5..5250c5a 100755 --- a/docs/apibuild.py +++ b/docs/apibuild.py @@ -926,7 +926,7 @@ class CParser: if i < len(l) and l[i] == ' ': i = i + 1 l = l[i:] - if len(l) >= 6 and l[0:7] == "returns" or l[0:7] == "Returns": + if len(l) >= 6 and l[0:7] == "Returns": try: l = string.split(l, ' ', 1)[1] except: -- 1.9.0

3 2