- Devel - libvirt List Archives

[libvirt] RFC decoupling VM NIC provisioning from VM NIC connection to backend networks
by Sumit Naiksatam (snaiksat) 02 Nov '11

02 Nov '11

Hi, In its current implementation Libvirt makes sure that the network interfaces that it passes/provision to a VM (for example to qemu[-kvm]) are already connected to its backend (interfaces/networks) by the time the VM starts its boot process. In a non virtualized setup it would be like booting a machine with the Ethernet cable already plugged into a router/switch port. While in a non virtualized setup you can boot a machine first (with no physical connection to a router/switch) and later connect its NIC/s to the switch/router, when you boot a VM via Libvirt it is not possible to decouple the two actions (VM boot, cable plug/unplug). An example of case where the capability of decoupling the two actions mentioned above is a requirement in Quantum/NetStack which is the network service leveraged by OpenStack. The modular design of OpenStack allows you to: - provision VMs with NIC/s - create networks - create ports on networks - plug/unplug a VM NIC into/from a given port on a network (at runtime) Note that this runtime plug/unplug requirement has nothing to do with hot plug/unplug of NICs. The idea is more that of decoupling the provisioning of a VM from the connection of the VM to the network/s. This would make it possible to change (at run-time too) the networks the NIC/s of a given VM are connected to. For example, when a VM boots, its interfaces should be in link down state if the network admin has not connected the VM NIC/s to any "network" yet. Even though libvirt already provides a way to change the link state of an a VM NIC, link state and physical connection are two different things and should be manageable independently. Ideally the configuration syntax should be interface type and hypervisor type agnostic. Let's take QEMU[-kvm] as an example - when Libvirt starts a QEMU VM, it passes to QEMU a number of file descriptors that map to host backend interfaces (for example macvtap interfaces). In order to introduce this runtime plug/unplug capability, we need a mechanism that permits to delay the binding between the host macvtap interfaces and the guest taps (because you cannot know the fd of the macvtap interfaces before you create them). This means you need a mechanism that allows you to change such fd/s at runtime: - you can close/reset an fd (ie, when you disconnect a VM NIC from its network) - you can open/set an fd (ie, when you connect a VM NIC to a network) This could probably be a libvirt command that translates to a QEMU monitor command. Can the runtime plug/unplug capability described above be achieved (cleanly) with another mechanism? Is anybody working on implementing something similar? [For more information on OpenStack/NetStack/Quantum and the above requirements please refer to the network model used therein: http://docs.openstack.org/incubation/openstack-network/admin/content/Wha tIsQuantum.html (information on network, port, and attachment abstractions) http://www.slideshare.net/danwent/quantum-diablo-summary (slides 7 & 8)] Thanks, ~Sumit Naiksatam. (On behalf of OpenStack/Quantum team)

4 6

[libvirt] [PATCHv2] esx: Support folders in the path of vpx:// connection URIs
by Matthias Bolte 02 Nov '11

02 Nov '11

Allow the datacenter and compute resource parts of the path to be prefixed with folders. Therefore, the way the path is parsed has changed. Before, it was split in 2 or 3 items and the items' meanings were determined by their positions. Now the path can have 2 or more items and the the vCenter server is asked whether a folder, datacenter of compute resource with the specified name exists at the current hierarchy level. Before the datacenter and compute resource lookup automatically traversed folders during lookup. This is logic got removed and folders have to be specified explicitly. The proper datacenter path including folders is now used when accessing a datastore over HTTPS. This makes virsh dumpxml and define work for datacenters in folders. https://bugzilla.redhat.com/show_bug.cgi?id=732676 --- v1: https://www.redhat.com/archives/libvir-list/2011-September/msg00345.html v2: - Fix managed object lookup for esx:// connections that was broken by the changes to esxVI_Context_LookupManagedObjectsByPath in v1 - Store parsed/generated path parts in the context docs/drvesx.html.in | 10 ++- src/esx/esx_driver.c | 30 +++-- src/esx/esx_util.c | 22 +--- src/esx/esx_util.h | 4 +- src/esx/esx_vi.c | 267 +++++++++++++++++++++++++++++++++++----- src/esx/esx_vi.h | 11 +- src/esx/esx_vi_generator.input | 4 + 7 files changed, 280 insertions(+), 68 deletions(-) diff --git a/docs/drvesx.html.in b/docs/drvesx.html.in index da9d2a1..7d323b3 100644 --- a/docs/drvesx.html.in +++ b/docs/drvesx.html.in @@ -56,7 +56,7 @@ esx://example-esx.com/?no_verify=1 (ESX over HTTPS, but doesn't verify the s URIs have this general form (<code>[...]</code> marks an optional part). </p> <pre> -type://[username@]hostname[:port]/[datacenter[/cluster]/server][?extraparameters] +type://[username@]hostname[:port]/[[folder/...]datacenter/[folder/...][cluster/]server][?extraparameters] </pre> <p> The <code>type://</code> is either <code>esx://</code> or @@ -80,6 +80,14 @@ type://[username@]hostname[:port]/[datacenter[/cluster]/server][?extraparameters <pre> vpx://example-vcenter.com/dc1/cluster1/example-esx.com </pre> + <p> + Datacenters and clusters can be organized in folders, those have to be + specified as well. The driver can handle folders + <span class="since">since 0.9.7</span>. + </p> +<pre> +vpx://example-vcenter.com/folder1/dc1/folder2/example-esx.com +</pre> <h4><a name="extraparams">Extra parameters</a></h4> diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index c5e4974..109d9f1 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -722,7 +722,7 @@ esxConnectToHost(esxPrivate *priv, virConnectAuthPtr auth, if (esxVI_Context_Alloc(&priv->host) < 0 || esxVI_Context_Connect(priv->host, url, ipAddress, username, password, priv->parsedUri) < 0 || - esxVI_Context_LookupObjectsByPath(priv->host, priv->parsedUri) < 0) { + esxVI_Context_LookupManagedObjects(priv->host) < 0) { goto cleanup; } @@ -799,8 +799,7 @@ esxConnectToVCenter(esxPrivate *priv, virConnectAuthPtr auth, char *url = NULL; if (hostSystemIpAddress == NULL && - (priv->parsedUri->path_datacenter == NULL || - priv->parsedUri->path_computeResource == NULL)) { + (priv->parsedUri->path == NULL || STREQ(priv->parsedUri->path, "/"))) { ESX_ERROR(VIR_ERR_INVALID_ARG, "%s", _("Path has to specify the datacenter and compute resource")); return -1; @@ -859,13 +858,13 @@ esxConnectToVCenter(esxPrivate *priv, virConnectAuthPtr auth, } if (hostSystemIpAddress != NULL) { - if (esxVI_Context_LookupObjectsByHostSystemIp(priv->vCenter, - hostSystemIpAddress) < 0) { + if (esxVI_Context_LookupManagedObjectsByHostSystemIp + (priv->vCenter, hostSystemIpAddress) < 0) { goto cleanup; } } else { - if (esxVI_Context_LookupObjectsByPath(priv->vCenter, - priv->parsedUri) < 0) { + if (esxVI_Context_LookupManagedObjectsByPath(priv->vCenter, + priv->parsedUri->path) < 0) { goto cleanup; } } @@ -884,8 +883,8 @@ esxConnectToVCenter(esxPrivate *priv, virConnectAuthPtr auth, /* - * URI format: {vpx|esx|gsx}://[<username>@]<hostname>[:<port>]/[<path>][?<query parameter> ...] - * <path> = <datacenter>/<computeresource>[/<hostsystem>] + * URI format: {vpx|esx|gsx}://[<username>@]<hostname>[:<port>]/[<path>][?<query parameter>...] + * <path> = [<folder>/...]<datacenter>/[<folder>/...]<computeresource>[/<hostsystem>] * * If no port is specified the default port is set dependent on the scheme and * transport parameter: @@ -899,7 +898,8 @@ esxConnectToVCenter(esxPrivate *priv, virConnectAuthPtr auth, * For a vpx:// connection <path> references a host managed by the vCenter. * In case the host is part of a cluster then <computeresource> is the cluster * name. Otherwise <computeresource> and <hostsystem> are equal and the later - * can be omitted. + * can be omitted. As datacenters and computeresources can be organized in + * folders those have to be included in <path>. * * Optional query parameters: * - transport={http|https} @@ -967,6 +967,12 @@ esxOpen(virConnectPtr conn, virConnectAuthPtr auth, return VIR_DRV_OPEN_ERROR; } + if (STRCASENEQ(conn->uri->scheme, "vpx") && + conn->uri->path != NULL && STRNEQ(conn->uri->path, "/")) { + VIR_WARN("Ignoring unexpected path '%s' for non-vpx scheme '%s'", + conn->uri->path, conn->uri->scheme); + } + /* Require server part */ if (conn->uri->server == NULL) { ESX_ERROR(VIR_ERR_INVALID_ARG, "%s", @@ -2759,7 +2765,7 @@ esxDomainGetXMLDesc(virDomainPtr domain, unsigned int flags) domain->conn->uri->server, domain->conn->uri->port); virBufferURIEncodeString(&buffer, directoryAndFileName); virBufferAddLit(&buffer, "?dcPath="); - virBufferURIEncodeString(&buffer, priv->primary->datacenter->name); + virBufferURIEncodeString(&buffer, priv->primary->datacenterPath); virBufferAddLit(&buffer, "&dsName="); virBufferURIEncodeString(&buffer, datastoreName); @@ -3227,7 +3233,7 @@ esxDomainDefineXML(virConnectPtr conn, const char *xml) virBufferURIEncodeString(&buffer, escapedName); virBufferAddLit(&buffer, ".vmx?dcPath="); - virBufferURIEncodeString(&buffer, priv->primary->datacenter->name); + virBufferURIEncodeString(&buffer, priv->primary->datacenterPath); virBufferAddLit(&buffer, "&dsName="); virBufferURIEncodeString(&buffer, datastoreName); diff --git a/src/esx/esx_util.c b/src/esx/esx_util.c index c14179d..1925802 100644 --- a/src/esx/esx_util.c +++ b/src/esx/esx_util.c @@ -51,7 +51,6 @@ esxUtil_ParseUri(esxUtil_ParsedUri **parsedUri, xmlURIPtr uri) int noVerify; int autoAnswer; char *tmp; - char *saveptr; if (parsedUri == NULL || *parsedUri != NULL) { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, "%s", _("Invalid argument")); @@ -184,26 +183,13 @@ esxUtil_ParseUri(esxUtil_ParsedUri **parsedUri, xmlURIPtr uri) } } - /* Expected format: [/]<datacenter>/<computeresource>[/<hostsystem>] */ if (uri->path != NULL) { - tmp = strdup(uri->path); + (*parsedUri)->path = strdup(uri->path); - if (tmp == NULL) { + if ((*parsedUri)->path == NULL) { virReportOOMError(); goto cleanup; } - - if (esxVI_String_DeepCopyValue(&(*parsedUri)->path_datacenter, - strtok_r(tmp, "/", &saveptr)) < 0 || - esxVI_String_DeepCopyValue(&(*parsedUri)->path_computeResource, - strtok_r(NULL, "/", &saveptr)) < 0 || - esxVI_String_DeepCopyValue(&(*parsedUri)->path_hostSystem, - strtok_r(NULL, "", &saveptr)) < 0) { - VIR_FREE(tmp); - goto cleanup; - } - - VIR_FREE(tmp); } if ((*parsedUri)->transport == NULL) { @@ -242,9 +228,7 @@ esxUtil_FreeParsedUri(esxUtil_ParsedUri **parsedUri) VIR_FREE((*parsedUri)->transport); VIR_FREE((*parsedUri)->vCenter); VIR_FREE((*parsedUri)->proxy_hostname); - VIR_FREE((*parsedUri)->path_datacenter); - VIR_FREE((*parsedUri)->path_computeResource); - VIR_FREE((*parsedUri)->path_hostSystem); + VIR_FREE((*parsedUri)->path); VIR_FREE(*parsedUri); } diff --git a/src/esx/esx_util.h b/src/esx/esx_util.h index 39fdb6d..8d172e1 100644 --- a/src/esx/esx_util.h +++ b/src/esx/esx_util.h @@ -37,9 +37,7 @@ struct _esxUtil_ParsedUri { int proxy_type; char *proxy_hostname; int proxy_port; - char *path_datacenter; - char *path_computeResource; - char *path_hostSystem; + char *path; }; int esxUtil_ParseUri(esxUtil_ParsedUri **parsedUri, xmlURIPtr uri); diff --git a/src/esx/esx_vi.c b/src/esx/esx_vi.c index 87ae802..314f322 100644 --- a/src/esx/esx_vi.c +++ b/src/esx/esx_vi.c @@ -616,8 +616,11 @@ ESX_VI__TEMPLATE__FREE(Context, esxVI_UserSession_Free(&item->session); VIR_FREE(item->sessionLock); esxVI_Datacenter_Free(&item->datacenter); + VIR_FREE(item->datacenterPath); esxVI_ComputeResource_Free(&item->computeResource); + VIR_FREE(item->computeResourcePath); esxVI_HostSystem_Free(&item->hostSystem); + VIR_FREE(item->hostSystemName); esxVI_SelectionSpec_Free(&item->selectSet_folderToChildEntity); esxVI_SelectionSpec_Free(&item->selectSet_hostSystemToParent); esxVI_SelectionSpec_Free(&item->selectSet_hostSystemToVm); @@ -789,21 +792,25 @@ esxVI_Context_Connect(esxVI_Context *ctx, const char *url, } int -esxVI_Context_LookupObjectsByPath(esxVI_Context *ctx, - esxUtil_ParsedUri *parsedUri) +esxVI_Context_LookupManagedObjects(esxVI_Context *ctx) { - char *hostSystemName = NULL; /* Lookup Datacenter */ - if (esxVI_LookupDatacenter(ctx, parsedUri->path_datacenter, - ctx->service->rootFolder, NULL, &ctx->datacenter, + if (esxVI_LookupDatacenter(ctx, NULL, ctx->service->rootFolder, NULL, + &ctx->datacenter, esxVI_Occurrence_RequiredItem) < 0) { return -1; } + ctx->datacenterPath = strdup(ctx->datacenter->name); + + if (ctx->datacenterPath == NULL) { + virReportOOMError(); + return -1; + } + /* Lookup (Cluster)ComputeResource */ - if (esxVI_LookupComputeResource(ctx, parsedUri->path_computeResource, - ctx->datacenter->hostFolder, NULL, - &ctx->computeResource, + if (esxVI_LookupComputeResource(ctx, NULL, ctx->datacenter->hostFolder, + NULL, &ctx->computeResource, esxVI_Occurrence_RequiredItem) < 0) { return -1; } @@ -814,38 +821,240 @@ esxVI_Context_LookupObjectsByPath(esxVI_Context *ctx, return -1; } + ctx->computeResourcePath = strdup(ctx->computeResource->name); + + if (ctx->computeResourcePath == NULL) { + virReportOOMError(); + return -1; + } + /* Lookup HostSystem */ - if (parsedUri->path_hostSystem == NULL && - STREQ(ctx->computeResource->_reference->type, - "ClusterComputeResource")) { - ESX_VI_ERROR(VIR_ERR_INVALID_ARG, "%s", - _("Path has to specify the host system")); + if (esxVI_LookupHostSystem(ctx, NULL, ctx->computeResource->_reference, + NULL, &ctx->hostSystem, + esxVI_Occurrence_RequiredItem) < 0) { return -1; } - if (parsedUri->path_hostSystem != NULL || - (parsedUri->path_computeResource != NULL && - parsedUri->path_hostSystem == NULL)) { - if (parsedUri->path_hostSystem != NULL) { - hostSystemName = parsedUri->path_hostSystem; + ctx->hostSystemName = strdup(ctx->hostSystem->name); + + if (ctx->hostSystemName == NULL) { + virReportOOMError(); + return -1; + } + + return 0; +} + +int +esxVI_Context_LookupManagedObjectsByPath(esxVI_Context *ctx, const char *path) +{ + int result = -1; + char *tmp = NULL; + char *saveptr = NULL; + char *previousItem = NULL; + char *item = NULL; + virBuffer buffer = VIR_BUFFER_INITIALIZER; + esxVI_ManagedObjectReference *root = NULL; + esxVI_Folder *folder = NULL; + + tmp = strdup(path); + + if (tmp == NULL) { + virReportOOMError(); + goto cleanup; + } + + /* Lookup Datacenter */ + item = strtok_r(tmp, "/", &saveptr); + + if (item == NULL) { + ESX_VI_ERROR(VIR_ERR_INVALID_ARG, + _("Path '%s' does not specify a datacenter"), path); + goto cleanup; + } + + root = ctx->service->rootFolder; + + while (ctx->datacenter == NULL && item != NULL) { + esxVI_Folder_Free(&folder); + + /* Try to lookup item as a folder */ + if (esxVI_LookupFolder(ctx, item, root, NULL, &folder, + esxVI_Occurrence_OptionalItem) < 0) { + goto cleanup; + } + + if (folder != NULL) { + /* It's a folder, use it as new lookup root */ + if (root != ctx->service->rootFolder) { + esxVI_ManagedObjectReference_Free(&root); + } + + root = folder->_reference; + folder->_reference = NULL; + } else { + /* Try to lookup item as a datacenter */ + if (esxVI_LookupDatacenter(ctx, item, root, NULL, &ctx->datacenter, + esxVI_Occurrence_OptionalItem) < 0) { + goto cleanup; + } + } + + /* Build path_datacenter */ + if (virBufferUse(&buffer) > 0) { + virBufferAddChar(&buffer, '/'); + } + + virBufferAdd(&buffer, item, -1); + + previousItem = item; + item = strtok_r(NULL, "/", &saveptr); + } + + if (ctx->datacenter == NULL) { + ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, + _("Could not find datacenter specified in '%s'"), path); + goto cleanup; + } + + if (virBufferError(&buffer)) { + virReportOOMError(); + goto cleanup; + } + + ctx->datacenterPath = virBufferContentAndReset(&buffer); + + /* Lookup (Cluster)ComputeResource */ + if (item == NULL) { + ESX_VI_ERROR(VIR_ERR_INVALID_ARG, + _("Path '%s' does not specify a compute resource"), path); + goto cleanup; + } + + if (root != ctx->service->rootFolder) { + esxVI_ManagedObjectReference_Free(&root); + } + + root = ctx->datacenter->hostFolder; + + while (ctx->computeResource == NULL && item != NULL) { + esxVI_Folder_Free(&folder); + + /* Try to lookup item as a folder */ + if (esxVI_LookupFolder(ctx, item, root, NULL, &folder, + esxVI_Occurrence_OptionalItem) < 0) { + goto cleanup; + } + + if (folder != NULL) { + /* It's a folder, use it as new lookup root */ + if (root != ctx->datacenter->hostFolder) { + esxVI_ManagedObjectReference_Free(&root); + } + + root = folder->_reference; + folder->_reference = NULL; } else { - hostSystemName = parsedUri->path_computeResource; + /* Try to lookup item as a compute resource */ + if (esxVI_LookupComputeResource(ctx, item, root, NULL, + &ctx->computeResource, + esxVI_Occurrence_OptionalItem) < 0) { + goto cleanup; + } + } + + /* Build path_computeResource */ + if (virBufferUse(&buffer) > 0) { + virBufferAddChar(&buffer, '/'); + } + + virBufferAdd(&buffer, item, -1); + + previousItem = item; + item = strtok_r(NULL, "/", &saveptr); + } + + if (ctx->computeResource == NULL) { + ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, + _("Could not find compute resource specified in '%s'"), + path); + goto cleanup; + } + + if (ctx->computeResource->resourcePool == NULL) { + ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, "%s", + _("Could not retrieve resource pool")); + goto cleanup; + } + + if (virBufferError(&buffer)) { + virReportOOMError(); + goto cleanup; + } + + ctx->computeResourcePath = virBufferContentAndReset(&buffer); + + /* Lookup HostSystem */ + if (STREQ(ctx->computeResource->_reference->type, + "ClusterComputeResource")) { + if (item == NULL) { + ESX_VI_ERROR(VIR_ERR_INVALID_ARG, + _("Path '%s' does not specify a host system"), path); + goto cleanup; } + + /* The path specified a cluster, it has to specify a host system too */ + previousItem = item; + item = strtok_r(NULL, "/", &saveptr); + } + + if (item != NULL) { + ESX_VI_ERROR(VIR_ERR_INVALID_ARG, + _("Path '%s' ends with an excess item"), path); + goto cleanup; } - if (esxVI_LookupHostSystem(ctx, hostSystemName, + ctx->hostSystemName = strdup(previousItem); + + if (ctx->hostSystemName == NULL) { + virReportOOMError(); + goto cleanup; + } + + if (esxVI_LookupHostSystem(ctx, ctx->hostSystemName, ctx->computeResource->_reference, NULL, &ctx->hostSystem, - esxVI_Occurrence_RequiredItem) < 0) { - return -1; + esxVI_Occurrence_OptionalItem) < 0) { + goto cleanup; } - return 0; + if (ctx->hostSystem == NULL) { + ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, + _("Could not find host system specified in '%s'"), path); + goto cleanup; + } + + result = 0; + + cleanup: + if (result < 0) { + virBufferFreeAndReset(&buffer); + } + + if (root != ctx->service->rootFolder && + (ctx->datacenter == NULL || root != ctx->datacenter->hostFolder)) { + esxVI_ManagedObjectReference_Free(&root); + } + + VIR_FREE(tmp); + esxVI_Folder_Free(&folder); + + return result; } int -esxVI_Context_LookupObjectsByHostSystemIp(esxVI_Context *ctx, - const char *hostSystemIpAddress) +esxVI_Context_LookupManagedObjectsByHostSystemIp(esxVI_Context *ctx, + const char *hostSystemIpAddress) { int result = -1; esxVI_ManagedObjectReference *managedObjectReference = NULL; @@ -1491,8 +1700,7 @@ esxVI_BuildSelectSetCollection(esxVI_Context *ctx) /* Folder -> childEntity (ManagedEntity) */ if (esxVI_BuildSelectSet(&ctx->selectSet_folderToChildEntity, "folderToChildEntity", - "Folder", "childEntity", - "folderToChildEntity\0") < 0) { + "Folder", "childEntity", NULL) < 0) { return -1; } @@ -1689,9 +1897,10 @@ esxVI_LookupObjectContentByType(esxVI_Context *ctx, objectSpec->obj = root; objectSpec->skip = esxVI_Boolean_False; - if (STRNEQ(root->type, type)) { + if (STRNEQ(root->type, type) || STREQ(root->type, "Folder")) { if (STREQ(root->type, "Folder")) { - if (STREQ(type, "Datacenter") || STREQ(type, "ComputeResource") || + if (STREQ(type, "Folder") || STREQ(type, "Datacenter") || + STREQ(type, "ComputeResource") || STREQ(type, "ClusterComputeResource")) { objectSpec->selectSet = ctx->selectSet_folderToChildEntity; } else { diff --git a/src/esx/esx_vi.h b/src/esx/esx_vi.h index 0b3f889..78d3986 100644 --- a/src/esx/esx_vi.h +++ b/src/esx/esx_vi.h @@ -204,8 +204,11 @@ struct _esxVI_Context { esxVI_UserSession *session; /* ... except the session ... */ virMutexPtr sessionLock; /* ... that is protected by this mutex */ esxVI_Datacenter *datacenter; + char *datacenterPath; /* including folders */ esxVI_ComputeResource *computeResource; + char *computeResourcePath; /* including folders */ esxVI_HostSystem *hostSystem; + char *hostSystemName; esxVI_SelectionSpec *selectSet_folderToChildEntity; esxVI_SelectionSpec *selectSet_hostSystemToParent; esxVI_SelectionSpec *selectSet_hostSystemToVm; @@ -221,10 +224,10 @@ void esxVI_Context_Free(esxVI_Context **ctx); int esxVI_Context_Connect(esxVI_Context *ctx, const char *ipAddress, const char *url, const char *username, const char *password, esxUtil_ParsedUri *parsedUri); -int esxVI_Context_LookupObjectsByPath(esxVI_Context *ctx, - esxUtil_ParsedUri *parsedUri); -int esxVI_Context_LookupObjectsByHostSystemIp(esxVI_Context *ctx, - const char *hostSystemIpAddress); +int esxVI_Context_LookupManagedObjects(esxVI_Context *ctx); +int esxVI_Context_LookupManagedObjectsByPath(esxVI_Context *ctx, const char *path); +int esxVI_Context_LookupManagedObjectsByHostSystemIp(esxVI_Context *ctx, + const char *hostSystemIpAddress); int esxVI_Context_Execute(esxVI_Context *ctx, const char *methodName, const char *request, esxVI_Response **response, esxVI_Occurrence occurrence); diff --git a/src/esx/esx_vi_generator.input b/src/esx/esx_vi_generator.input index 361a6e7..1a67a8c 100644 --- a/src/esx/esx_vi_generator.input +++ b/src/esx/esx_vi_generator.input @@ -755,6 +755,10 @@ managed object Datacenter extends ManagedEntity end +managed object Folder extends ManagedEntity +end + + managed object HostSystem extends ManagedEntity HostConfigManager configManager r end -- 1.7.4.1

2 2

[libvirt] [PATCH] esx: Support vSphere 5.x
by Patrice LACHANCE 02 Nov '11

02 Nov '11

Hello Based on http://www.redhat.com/archives/libvir-list/2010-July/msg00480.html, I created this quick patch to accept connection to vSphere 5.x Notes: - I had to remove the warnings for version > 4.1 & 5 because they were generating errors when running scripts encapsulating virsh commands with opennebula. A better option might be to relie on an environment variable (for example DONT_BLAME_LIBVIRT!) to toggle warnings on/off - Ran only a few tests but it works as expected. - Next steps should be to work on VMware API 5. integration Patrice diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index 41086ef..78872d4 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -730,9 +730,11 @@ esxConnectToHost(esxPrivate *priv, virConnectAuthPtr auth, if (priv->host->productVersion != esxVI_ProductVersion_ESX35 && priv->host->productVersion != esxVI_ProductVersion_ESX40 && priv->host->productVersion != esxVI_ProductVersion_ESX41 && - priv->host->productVersion != esxVI_ProductVersion_ESX4x) { + priv->host->productVersion != esxVI_ProductVersion_ESX4x && + priv->host->productVersion != esxVI_ProductVersion_ESX50 && + priv->host->productVersion != esxVI_ProductVersion_ESX5x) { ESX_ERROR(VIR_ERR_INTERNAL_ERROR, - _("%s is neither an ESX 3.5 host nor an ESX 4.x host"), + _("%s is neither an ESX 3.5, 4.x nor 5.x host"), hostname); goto cleanup; } @@ -857,10 +859,11 @@ esxConnectToVCenter(esxPrivate *priv, virConnectAuthPtr auth, if (priv->vCenter->productVersion != esxVI_ProductVersion_VPX25 && priv->vCenter->productVersion != esxVI_ProductVersion_VPX40 && priv->vCenter->productVersion != esxVI_ProductVersion_VPX41 && - priv->vCenter->productVersion != esxVI_ProductVersion_VPX4x) { + priv->vCenter->productVersion != esxVI_ProductVersion_VPX4x && + priv->vCenter->productVersion != esxVI_ProductVersion_VPX50 && + priv->vCenter->productVersion != esxVI_ProductVersion_VPX5x) { ESX_ERROR(VIR_ERR_INTERNAL_ERROR, - _("%s is neither a vCenter 2.5 server nor a vCenter " - "4.x server"), hostname); + _("%s is neither a vCenter 2.5, 4.x nor 5.x server"), hostname); goto cleanup; } diff --git a/src/esx/esx_vi.c b/src/esx/esx_vi.c index 325157c..3a11116 100644 --- a/src/esx/esx_vi.c +++ b/src/esx/esx_vi.c @@ -675,9 +675,23 @@ esxVI_Context_Connect(esxVI_Context *ctx, const char *url, VIR_WARN("Found untested VI API major/minor version '%s'", ctx->service->about->apiVersion); + } else if (STRPREFIX(ctx->service->about->apiVersion, "5.0")) { + ctx->apiVersion = esxVI_APIVersion_50; + + /* + VIR_WARN("Found untested VI API major/minor version '%s'", + ctx->service->about->apiVersion); + */ + } else if (STRPREFIX(ctx->service->about->apiVersion, "5.")) { + ctx->apiVersion = esxVI_APIVersion_5x; + + /* + VIR_WARN("Found untested VI API major/minor version '%s'", + ctx->service->about->apiVersion); + */ } else { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, - _("Expecting VI API major/minor version '2.5' or '4.x' " + _("Expecting VI API major/minor version '2.5', '4.x' or '5.x' " "but found '%s'"), ctx->service->about->apiVersion); return -1; } @@ -704,10 +718,24 @@ esxVI_Context_Connect(esxVI_Context *ctx, const char *url, VIR_WARN("Found untested ESX major/minor version '%s'", ctx->service->about->version); + } else if (STRPREFIX(ctx->service->about->version, "5.0")) { + ctx->productVersion = esxVI_ProductVersion_ESX50; + + /* + VIR_WARN("Found untested ESX major/minor version '%s'", + ctx->service->about->version); + */ + } else if (STRPREFIX(ctx->service->about->version, "5.")) { + ctx->productVersion = esxVI_ProductVersion_ESX5x; + + /* + VIR_WARN("Found untested ESX major/minor version '%s'", + ctx->service->about->version); + */ } else { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, - _("Expecting ESX major/minor version '3.5' or " - "'4.x' but found '%s'"), + _("Expecting ESX major/minor version '3.5', '4.x' or " + "'5.x' but found '%s'"), ctx->service->about->version); return -1; } @@ -723,9 +751,23 @@ esxVI_Context_Connect(esxVI_Context *ctx, const char *url, VIR_WARN("Found untested VPX major/minor version '%s'", ctx->service->about->version); + } else if (STRPREFIX(ctx->service->about->version, "5.0")) { + ctx->productVersion = esxVI_ProductVersion_VPX50; + + /* + VIR_WARN("Found untested VPX major/minor version '%s'", + ctx->service->about->version); + */ + } else if (STRPREFIX(ctx->service->about->version, "5.")) { + ctx->productVersion = esxVI_ProductVersion_VPX5x; + + /* + VIR_WARN("Found untested VPX major/minor version '%s'", + ctx->service->about->version); + */ } else { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, - _("Expecting VPX major/minor version '2.5' or '4.x' " + _("Expecting VPX major/minor version '2.5', '4.x' or '5.x' " "but found '%s'"), ctx->service->about->version); return -1; } @@ -3919,6 +3961,12 @@ esxVI_ProductVersionToDefaultVirtualHWVersion(esxVI_ProductVersion productVersio case esxVI_ProductVersion_VPX4x: return 7; + case esxVI_ProductVersion_ESX50: + case esxVI_ProductVersion_VPX50: + case esxVI_ProductVersion_ESX5x: + case esxVI_ProductVersion_VPX5x: + return 8; + default: ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, "%s", _("Unexpected product version")); diff --git a/src/esx/esx_vi.h b/src/esx/esx_vi.h index b8e921f..fbf3fb2 100644 --- a/src/esx/esx_vi.h +++ b/src/esx/esx_vi.h @@ -99,7 +99,9 @@ enum _esxVI_APIVersion { esxVI_APIVersion_25, esxVI_APIVersion_40, esxVI_APIVersion_41, - esxVI_APIVersion_4x /* > 4.1 */ + esxVI_APIVersion_4x, + esxVI_APIVersion_50, + esxVI_APIVersion_5x /* > 5.0 */ }; /* @@ -116,13 +118,17 @@ enum _esxVI_ProductVersion { esxVI_ProductVersion_ESX35 = esxVI_ProductVersion_ESX | 1, esxVI_ProductVersion_ESX40 = esxVI_ProductVersion_ESX | 2, esxVI_ProductVersion_ESX41 = esxVI_ProductVersion_ESX | 3, - esxVI_ProductVersion_ESX4x = esxVI_ProductVersion_ESX | 4, /* > 4.1 */ + esxVI_ProductVersion_ESX4x = esxVI_ProductVersion_ESX | 4, + esxVI_ProductVersion_ESX50 = esxVI_ProductVersion_ESX | 5, + esxVI_ProductVersion_ESX5x = esxVI_ProductVersion_ESX | 6, /* > 5.0 */ esxVI_ProductVersion_VPX = (1 << 2) << 16, esxVI_ProductVersion_VPX25 = esxVI_ProductVersion_VPX | 1, esxVI_ProductVersion_VPX40 = esxVI_ProductVersion_VPX | 2, esxVI_ProductVersion_VPX41 = esxVI_ProductVersion_VPX | 3, - esxVI_ProductVersion_VPX4x = esxVI_ProductVersion_VPX | 4 /* > 4.1 */ + esxVI_ProductVersion_VPX4x = esxVI_ProductVersion_VPX | 4, + esxVI_ProductVersion_VPX50 = esxVI_ProductVersion_VPX | 5, + esxVI_ProductVersion_VPX5x = esxVI_ProductVersion_VPX | 6 /* > 5.0 */ }; enum _esxVI_Occurrence {

2 1

[libvirt] [PATCH v3] Fix URI alias prefix matching
by Wen Ruo Lv 02 Nov '11

02 Nov '11

with /etc/libvirt/libvirt.conf below: uri_aliases = [ "hail=qemu:///system", "sleet=qemu+ssh://root 9 115 122 57/system", "sam=qemu+unix:///system?socket=/var/run/libvirt/libvirt-sock", ] Neither "virsh -c hailly" nor "hai" should result in matching "hail=qemu:///system" Fix URI alias prefix matching when connecting Signed-off-by: Wen Ruo Lv <lvroyce(a)linux.vnet.ibm.com> --- src/libvirt.c | 6 +++++- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index e9d1a29..6f8a76f 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -990,6 +990,8 @@ static int virConnectOpenFindURIAliasMatch(virConfValuePtr value, const char *alias, char **uri) { virConfValuePtr entry; + size_t alias_len; + if (value->type != VIR_CONF_LIST) { virLibConnError(VIR_ERR_INTERNAL_ERROR, "%s", _("Expected a list for 'uri_aliases' config parameter")); @@ -997,6 +999,7 @@ virConnectOpenFindURIAliasMatch(virConfValuePtr value, const char *alias, char * } entry = value->list; + alias_len = strlen(alias); while (entry) { char *offset; size_t safe; @@ -1022,7 +1025,8 @@ virConnectOpenFindURIAliasMatch(virConfValuePtr value, const char *alias, char * return -1; } - if (STREQLEN(entry->str, alias, offset-entry->str)) { + if (alias_len == (offset - entry->str) && + STREQLEN(entry->str, alias, alias_len)) { VIR_DEBUG("Resolved alias '%s' to '%s'", alias, offset+1); if (!(*uri = strdup(offset+1))) { -- 1.7.4.1

3 2

Re: [libvirt] [libvirt-users] Error on the site
by Eric Blake 02 Nov '11

02 Nov '11

On 11/01/2011 01:15 AM, Alexander Biryukov wrote: > Good afternoon. > I apologize for my English. > > Error on page: http://libvirt.org/formatdomain.html > To be more precise in the line "<target *def*='hdc' bus='ide'/>" Thanks; I'm pushing this as the trivial fix: diff --git i/docs/formatdomain.html.in w/docs/formatdomain.html.in index 4fb5829..c3e7752 100644 --- i/docs/formatdomain.html.in +++ w/docs/formatdomain.html.in @@ -925,7 +925,7 @@ </disk> <disk type='block' device='cdrom'> <driver name='qemu' type='raw'/> - <target def='hdc' bus='ide'/> + <target dev='hdc' bus='ide'/> <readonly/> </disk> </devices> -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 0

[libvirt] [PATCH] Don't overwrite error message during VM cleanup
by Daniel P. Berrange 02 Nov '11

02 Nov '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> If an LXC VM fails to start, quite a few cleanup paths will result in the original error message being overwritten. Some other cleanup paths also forgot to actually terminate the VM. * src/lxc/lxc_driver.c: Ensure VM is terminated on startup failure and preserve original error --- src/lxc/lxc_driver.c | 29 +++++++++++++++++++++-------- 1 files changed, 21 insertions(+), 8 deletions(-) diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c index 23efe1e..eabc1fb 100644 --- a/src/lxc/lxc_driver.c +++ b/src/lxc/lxc_driver.c @@ -1622,6 +1622,7 @@ static int lxcVmStart(virConnectPtr conn, char *timestamp; virCommandPtr cmd = NULL; lxcDomainObjPrivatePtr priv = vm->privateData; + virErrorPtr err = NULL; if (!lxc_driver->cgroup) { lxcError(VIR_ERR_INTERNAL_ERROR, "%s", @@ -1753,8 +1754,7 @@ static int lxcVmStart(virConnectPtr conn, _("guest failed to start: %s"), out); } - lxcVmTerminate(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED); - goto cleanup; + goto error; } if ((priv->monitorWatch = virEventAddHandle( @@ -1762,31 +1762,33 @@ static int lxcVmStart(virConnectPtr conn, VIR_EVENT_HANDLE_ERROR | VIR_EVENT_HANDLE_HANGUP, lxcMonitorEvent, vm, NULL)) < 0) { - lxcVmTerminate(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED); - goto cleanup; + goto error; } if (autoDestroy && lxcProcessAutoDestroyAdd(driver, vm, conn) < 0) - goto cleanup; + goto error; /* * Again, need to save the live configuration, because the function * requires vm->def->id != -1 to save tty info surely. */ if (virDomainSaveConfig(driver->stateDir, vm->def) < 0) - goto cleanup; + goto error; if (virDomainObjSetDefTransient(driver->caps, vm, false) < 0) - goto cleanup; + goto error; /* Write domain status to disk. */ if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) - goto cleanup; + goto error; rc = 0; cleanup: + if (rc != 0 && !err) + err = virSaveLastError(); + VIR_WARN("Cleanup %s" , err && err->message ? err->message : NULL); virCommandFree(cmd); if (VIR_CLOSE(logfd) < 0) { virReportSystemError(errno, "%s", _("could not close logfile")); @@ -1805,7 +1807,18 @@ cleanup: VIR_FORCE_CLOSE(handshakefds[0]); VIR_FORCE_CLOSE(handshakefds[1]); VIR_FREE(logfile); + + if (err) { + virSetError(err); + virResetError(err); + } + return rc; + +error: + err = virSaveLastError(); + lxcVmTerminate(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED); + goto cleanup; } /** -- 1.7.6.4

2 1

[libvirt] [PATCH] Add support for probing filesystem with libblkid
by Daniel P. Berrange 02 Nov '11

02 Nov '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> The LXC code for mounting container filesystems from block devices tries all filesystems in /etc/filesystems and possibly those in /proc/filesystems. The regular mount binary, however, first tries using libblkid to detect the format. Add support for doing the same in libvirt, since Fedora's /etc/filesystems is missing many formats, most notably ext4 which is the default filesystem Fedora uses! * src/Makefile.am: Link libvirt_lxc to libblkid * src/lxc/lxc_container.c: Probe filesystem format with liblkid --- src/Makefile.am | 4 ++ src/lxc/lxc_container.c | 102 ++++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 105 insertions(+), 1 deletions(-) diff --git a/src/Makefile.am b/src/Makefile.am index c36664b..bf26b19 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -1495,6 +1495,10 @@ libvirt_lxc_CFLAGS = \ $(AUDIT_CFLAGS) \ -I@top_srcdir@/src/conf \ $(AM_CFLAGS) +if HAVE_LIBBLKID +libvirt_lxc_CFLAGS += $(BLKID_CFLAGS) +libvirt_lxc_LDADD += $(BLKID_LIBS) +endif endif endif EXTRA_DIST += $(LXC_CONTROLLER_SOURCES) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index 36ac3a9..647aaea 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -47,6 +47,10 @@ # include <cap-ng.h> #endif +#if HAVE_LIBBLKID +# include <blkid/blkid.h> +#endif + #include "virterror_internal.h" #include "logging.h" #include "lxc_container.h" @@ -627,6 +631,86 @@ cleanup: } +#ifdef HAVE_LIBBLKID +static int +lxcContainerMountDetectFilesystem(const char *src, char **type) +{ + int fd; + int ret = -1; + int rc; + const char *data = NULL; + blkid_probe blkid = NULL; + + *type = NULL; + + if ((fd = open(src, O_RDONLY)) < 0) { + virReportSystemError(errno, + _("Unable to open filesystem %s"), src); + return -1; + } + + if (!(blkid = blkid_new_probe())) { + virReportSystemError(errno, "%s", + _("Unable to create blkid library handle")); + goto cleanup; + } + if (blkid_probe_set_device(blkid, fd, 0, 0) < 0) { + virReportSystemError(EINVAL, + _("Unable to associated device %s with blkid library"), + src); + goto cleanup; + } + + blkid_probe_enable_superblocks(blkid, 1); + + blkid_probe_set_superblocks_flags(blkid, BLKID_SUBLKS_TYPE); + + rc = blkid_do_safeprobe(blkid); + if (rc != 0) { + if (rc == 1) /* Nothing found, return success with *type == NULL */ + goto done; + + if (rc == -2) { + virReportSystemError(EINVAL, + _("Too many filesystems detected for %s"), + src); + } else { + virReportSystemError(errno, + _("Unable to detect filesystem for %s"), + src); + } + goto cleanup; + } + + if (blkid_probe_lookup_value(blkid, "TYPE", &data, NULL) < 0) { + virReportSystemError(ENOENT, + _("Unable to find filesystem type for %s"), + src); + goto cleanup; + } + + if (!(*type = strdup(data))) { + virReportOOMError(); + goto cleanup; + } + +done: + ret = 0; +cleanup: + VIR_FORCE_CLOSE(fd); + blkid_free_probe(blkid); + return ret; +} +#else /* ! HAVE_LIBBLKID */ +static int +lxcContainerMountDetectFilesystem(const char *src ATTRIBUTE_UNUSED, + char **type) +{ + /* No libblkid, so just return success with no detected type */ + *type = NULL; + return 0; +} +#endif /* ! HAVE_LIBBLKID */ /* * This functions attempts to do automatic detection of filesystem @@ -771,6 +855,8 @@ static int lxcContainerMountFSBlockHelper(virDomainFSDefPtr fs, { int fsflags = 0; int ret = -1; + char *format = NULL; + if (fs->readonly) fsflags |= MS_RDONLY; @@ -781,7 +867,21 @@ static int lxcContainerMountFSBlockHelper(virDomainFSDefPtr fs, goto cleanup; } - ret = lxcContainerMountFSBlockAuto(fs, fsflags, src, srcprefix); + if (lxcContainerMountDetectFilesystem(src, &format) < 0) + goto cleanup; + + if (format) { + VIR_DEBUG("Mount %s with detected format %s", src, format); + if (mount(src, fs->dst, format, fsflags, NULL) < 0) { + virReportSystemError(errno, + _("Failed to mount device %s to %s as %s"), + src, fs->dst, format); + goto cleanup; + } + ret = 0; + } else { + ret = lxcContainerMountFSBlockAuto(fs, fsflags, src, srcprefix); + } cleanup: return ret; -- 1.7.6.4

2 2

[libvirt] [PATCH] Fix error message when failing to detect filesystem
by Daniel P. Berrange 02 Nov '11

02 Nov '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> If we looped through /etc/filesystems trying to mount with each type and failed all options, we forget to actually raise an error message. * src/lxc/lxc_container.c: Raise error if unable to detect the filesystems. Also fix existing error message --- src/lxc/lxc_container.c | 8 +++++++- 1 files changed, 7 insertions(+), 1 deletions(-) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index fbd5267..36ac3a9 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -725,7 +725,7 @@ retry: continue; virReportSystemError(errno, - _("Failed to bind mount directory %s to %s"), + _("Failed to mount device %s to %s"), src, fs->dst); goto cleanup; } @@ -746,6 +746,12 @@ retry: goto retry; } + if (ret != 0) { + virReportSystemError(ENODEV, + _("Failed to mount device %s to %s, unable to detect filesystem"), + src, fs->dst); + } + VIR_DEBUG("Done mounting filesystem ret=%d tryProc=%d", ret, tryProc); cleanup: -- 1.7.6.4

3 2

[libvirt] [PATCH] ServerClient: Flush SASL data
by Michal Privoznik 02 Nov '11

02 Nov '11

If daemon is using SASL it reads client data into a cache. This cache is big (usually 65KB) and can thus contain 2 or more messages. However, on socket event we can dispatch only one message. So if we read two messages at once, the second will not be dispatched as the socket event goes away with filling the cache. Moreover, when dispatching the cache we need to remember to take care of client max requests limit. --- src/rpc/virnetserverclient.c | 34 +++++++++++++++++++++++++++++++++- 1 files changed, 33 insertions(+), 1 deletions(-) diff --git a/src/rpc/virnetserverclient.c b/src/rpc/virnetserverclient.c index 1256f0f..69af746 100644 --- a/src/rpc/virnetserverclient.c +++ b/src/rpc/virnetserverclient.c @@ -71,6 +71,8 @@ struct _virNetServerClient virNetTLSSessionPtr tls; #if HAVE_SASL virNetSASLSessionPtr sasl; + int sasl_timer; /* Timer to be fired upon cached data, + so we jump out from poll() immediately */ #endif /* Count of messages in the 'tx' queue, @@ -103,6 +105,7 @@ struct _virNetServerClient static void virNetServerClientDispatchEvent(virNetSocketPtr sock, int events, void *opaque); static void virNetServerClientUpdateEvent(virNetServerClientPtr client); +static void virNetServerClientDispatchRead(virNetServerClientPtr client); static void virNetServerClientLock(virNetServerClientPtr client) { @@ -191,6 +194,19 @@ static int virNetServerClientRegisterEvent(virNetServerClientPtr client) return 0; } +static void virNetServerClientDispatchReadTimerFunc(int timer, + void *opaque) +{ + virNetServerClientPtr client = opaque; + virNetServerClientLock(client); + virEventUpdateTimeout(timer, -1); + /* Even if this timer is enabled iff client->rx != NULL, + * but meanwhile client was unlocked, so it might have changed */ + if (client->rx) + virNetServerClientDispatchRead(client); + virNetServerClientUnlock(client); +} + /* * @client: a locked client object */ @@ -204,6 +220,19 @@ static void virNetServerClientUpdateEvent(virNetServerClientPtr client) mode = virNetServerClientCalculateHandleMode(client); virNetSocketUpdateIOCallback(client->sock, mode); +#ifdef HAVE_SASL + if (client->nrequests < client->nrequests_max && + client->rx && + virNetSocketHasCachedData(client->sock)) { + if (client->sasl_timer) + virEventUpdateTimeout(client->sasl_timer, 0); + else + client->sasl_timer = virEventAddTimeout(0, + virNetServerClientDispatchReadTimerFunc, + client, + NULL); + } +#endif } @@ -556,6 +585,8 @@ void virNetServerClientFree(virNetServerClientPtr client) VIR_FREE(client->identity); #if HAVE_SASL virNetSASLSessionFree(client->sasl); + if (client->sasl_timer) + virEventRemoveTimeout(client->sasl_timer); #endif virNetTLSSessionFree(client->tls); virNetTLSContextFree(client->tlsCtxt); @@ -990,7 +1021,8 @@ virNetServerClientDispatchEvent(virNetSocketPtr sock, int events, void *opaque) } else { if (events & VIR_EVENT_HANDLE_WRITABLE) virNetServerClientDispatchWrite(client); - if (events & VIR_EVENT_HANDLE_READABLE) + if (events & VIR_EVENT_HANDLE_READABLE && + client->rx) virNetServerClientDispatchRead(client); } } -- 1.7.3.4

3 5

[libvirt] [PATCH] Workaround for broken kernel autofs mounts
by Daniel P. Berrange 02 Nov '11

02 Nov '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> The kernel automounter is mostly broken wrt to containers. Most notably if you start a new filesystem namespace and then attempt to unmount any autofs filesystem, it will typically fail with a wierd error message like Failed to unmount '/.oldroot/sys/kernel/security':Too many levels of symbolic links Atttempting to detach the autofs mount mount using umount2(MNT_DETACH) will also fail with the same error. Therefore if we get any error on unmount()ing a filesystem from the old root FS when starting a container, we must immediately break out and detach the entire old root filesystem (ignoring any mounts below it). This has the effect of making the old root filesystem inaccessible to anything inside the container, but at the cost that the mounts live on in the kernel until the container exits. Given that SystemD uses autofs by default, we need LXC to be robust this scenario and thus this tradeoff is worthwhile. * src/lxc/lxc_container.c: Detach root filesystem if any umount operation fails. --- src/lxc/lxc_container.c | 49 ++++++++++++++++++++++++++++++++++++---------- 1 files changed, 38 insertions(+), 11 deletions(-) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index 63284e5..fbd5267 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -860,6 +860,9 @@ static int lxcContainerUnmountOldFS(void) FILE *procmnt; int i; char mntbuf[1024]; + int saveErrno; + const char *failedUmount = NULL; + int ret = -1; if (!(procmnt = setmntent("/proc/mounts", "r"))) { virReportSystemError(errno, "%s", @@ -872,17 +875,15 @@ static int lxcContainerUnmountOldFS(void) continue; if (VIR_REALLOC_N(mounts, nmounts+1) < 0) { - endmntent(procmnt); virReportOOMError(); - return -1; + goto cleanup; } if (!(mounts[nmounts++] = strdup(mntent.mnt_dir))) { - endmntent(procmnt); virReportOOMError(); - return -1; + goto cleanup; } + VIR_DEBUG("Grabbed %s", mntent.mnt_dir); } - endmntent(procmnt); if (mounts) qsort(mounts, nmounts, sizeof(mounts[0]), @@ -891,16 +892,42 @@ static int lxcContainerUnmountOldFS(void) for (i = 0 ; i < nmounts ; i++) { VIR_DEBUG("Umount %s", mounts[i]); if (umount(mounts[i]) < 0) { - virReportSystemError(errno, - _("Failed to unmount '%s'"), - mounts[i]); - return -1; + char ebuf[1024]; + failedUmount = mounts[i]; + saveErrno = errno; + VIR_WARN(_("Failed to unmount '%s', trying to detach root '%s': %s"), + failedUmount, mounts[nmounts-1], + virStrerror(errno, ebuf, sizeof(ebuf))); + break; } - VIR_FREE(mounts[i]); } + + if (failedUmount) { + /* This detaches the old root filesystem */ + if (umount2(mounts[nmounts-1], MNT_DETACH) < 0) { + virReportSystemError(saveErrno, + _("Failed to unmount '%s' and could not detach old root '%s'"), + failedUmount, mounts[nmounts-1]); + goto cleanup; + } + /* This unmounts the tmpfs on which the old root filesystem was hosted */ + if (umount(mounts[nmounts-1]) < 0) { + virReportSystemError(saveErrno, + _("Failed to unmount '%s' and could not unmount old root '%s'"), + failedUmount, mounts[nmounts-1]); + goto cleanup; + } + } + + ret = 0; + +cleanup: + for (i = 0 ; i < nmounts ; i++) + VIR_FREE(mounts[i]); + endmntent(procmnt); VIR_FREE(mounts); - return 0; + return ret; } -- 1.7.6.4

2 1