use of com.cloud.legacymodel.exceptions.CloudRuntimeException in project cosmic by MissionCriticalCloud.
the class VirtualMachineManagerImpl method advanceExpunge.
protected void advanceExpunge(VMInstanceVO vm) throws ResourceUnavailableException, OperationTimedoutException, ConcurrentOperationException {
if (vm == null || vm.getRemoved() != null) {
s_logger.debug("Unable to find vm or vm is destroyed: " + vm);
return;
}
advanceStop(vm.getUuid(), false);
vm = _vmDao.findByUuid(vm.getUuid());
try {
if (!stateTransitTo(vm, VirtualMachine.Event.ExpungeOperation, vm.getHostId())) {
s_logger.debug("Unable to destroy the vm because it is not in the correct state: " + vm);
throw new CloudRuntimeException("Unable to destroy " + vm);
}
} catch (final NoTransitionException e) {
s_logger.debug("Unable to destroy the vm because it is not in the correct state: " + vm);
throw new CloudRuntimeException("Unable to destroy " + vm, e);
}
s_logger.debug("Destroying vm " + vm);
final VirtualMachineProfile profile = new VirtualMachineProfileImpl(vm);
final HypervisorGuru hvGuru = _hvGuruMgr.getGuru(vm.getHypervisorType());
s_logger.debug("Cleaning up NICS");
final List<Command> nicExpungeCommands = hvGuru.finalizeExpungeNics(vm, profile.getNics());
_networkMgr.cleanupNics(profile);
s_logger.debug("Cleaning up hypervisor data structures (ex. SRs in XenServer) for managed storage");
final List<Command> volumeExpungeCommands = hvGuru.finalizeExpungeVolumes(vm);
final Long hostId = vm.getHostId() != null ? vm.getHostId() : vm.getLastHostId();
if (volumeExpungeCommands != null && volumeExpungeCommands.size() > 0 && hostId != null) {
final Commands cmds = new Commands(Command.OnError.Stop);
for (final Command volumeExpungeCommand : volumeExpungeCommands) {
cmds.addCommand(volumeExpungeCommand);
}
_agentMgr.send(hostId, cmds);
if (!cmds.isSuccessful()) {
for (final Answer answer : cmds.getAnswers()) {
if (!answer.getResult()) {
s_logger.warn("Failed to expunge vm due to: " + answer.getDetails());
throw new CloudRuntimeException("Unable to expunge " + vm + " due to " + answer.getDetails());
}
}
}
}
if (hostId != null) {
volumeMgr.revokeAccess(vm.getId(), hostId);
}
// Clean up volumes based on the vm's instance id
volumeMgr.cleanupVolumes(vm.getId());
final VirtualMachineGuru guru = getVmGuru(vm);
guru.finalizeExpunge(vm);
// remove the overcommit detials from the uservm details
_uservmDetailsDao.removeDetails(vm.getId());
// send hypervisor-dependent commands before removing
final List<Command> finalizeExpungeCommands = hvGuru.finalizeExpunge(vm);
if (finalizeExpungeCommands != null && finalizeExpungeCommands.size() > 0) {
if (hostId != null) {
final Commands cmds = new Commands(Command.OnError.Stop);
for (final Command command : finalizeExpungeCommands) {
cmds.addCommand(command);
}
if (nicExpungeCommands != null) {
for (final Command command : nicExpungeCommands) {
cmds.addCommand(command);
}
}
_agentMgr.send(hostId, cmds);
if (!cmds.isSuccessful()) {
for (final Answer answer : cmds.getAnswers()) {
if (!answer.getResult()) {
s_logger.warn("Failed to expunge vm due to: " + answer.getDetails());
throw new CloudRuntimeException("Unable to expunge " + vm + " due to " + answer.getDetails());
}
}
}
}
}
s_logger.debug("Expunged " + vm);
}
use of com.cloud.legacymodel.exceptions.CloudRuntimeException in project cosmic by MissionCriticalCloud.
the class VirtualMachineManagerImpl method orchestrateStart.
@Override
public void orchestrateStart(final String vmUuid, final Map<VirtualMachineProfile.Param, Object> params, final DeploymentPlan planToDeploy, final DeploymentPlanner planner) throws InsufficientCapacityException, ConcurrentOperationException, ResourceUnavailableException {
final CallContext cctxt = CallContext.current();
final Account account = cctxt.getCallingAccount();
final User caller = cctxt.getCallingUser();
VMInstanceVO vm = _vmDao.findByUuid(vmUuid);
final Ternary<VMInstanceVO, ReservationContext, ItWorkVO> start = changeToStartState(vm, caller, account);
if (start == null) {
return;
}
vm = start.first();
final ReservationContext ctx = start.second();
ItWorkVO work = start.third();
VMInstanceVO startedVm = null;
final ServiceOfferingVO offering = _offeringDao.findById(vm.getId(), vm.getServiceOfferingId());
final VirtualMachineTemplate template = _entityMgr.findByIdIncludingRemoved(VirtualMachineTemplate.class, vm.getTemplateId());
s_logger.debug("Trying to deploy VM, vm has dcId: " + vm.getDataCenterId() + " and podId: " + vm.getPodIdToDeployIn());
DataCenterDeployment plan = new DataCenterDeployment(vm.getDataCenterId(), vm.getPodIdToDeployIn(), null, null, null, null, ctx);
if (planToDeploy != null && planToDeploy.getDataCenterId() != 0) {
s_logger.debug("advanceStart: DeploymentPlan is provided, using dcId:" + planToDeploy.getDataCenterId() + ", podId: " + planToDeploy.getPodId() + ", clusterId: " + planToDeploy.getClusterId() + ", hostId: " + planToDeploy.getHostId() + ", poolId: " + planToDeploy.getPoolId());
plan = new DataCenterDeployment(planToDeploy.getDataCenterId(), planToDeploy.getPodId(), planToDeploy.getClusterId(), planToDeploy.getHostId(), planToDeploy.getPoolId(), planToDeploy.getPhysicalNetworkId(), ctx);
}
final HypervisorGuru hvGuru = _hvGuruMgr.getGuru(vm.getHypervisorType());
final VirtualMachineGuru vmGuru = getVmGuru(vm);
boolean canRetry = true;
ExcludeList avoids = null;
try {
final Journal journal = start.second().getJournal();
if (planToDeploy != null) {
avoids = planToDeploy.getAvoids();
}
if (avoids == null) {
avoids = new ExcludeList();
}
s_logger.debug("VM start orchestration will {}", avoids.toString());
boolean planChangedByVolume = false;
boolean reuseVolume = true;
final DataCenterDeployment originalPlan = plan;
int retry = StartRetry.value();
while (retry-- != 0) {
if (reuseVolume) {
// edit plan if this vm's ROOT volume is in READY state already
final List<VolumeVO> vols = _volsDao.findReadyRootVolumesByInstance(vm.getId());
for (final VolumeVO vol : vols) {
// make sure if the templateId is unchanged. If it is changed,
// let planner
// reassign pool for the volume even if it ready.
final Long volTemplateId = vol.getTemplateId();
if (volTemplateId != null && volTemplateId != template.getId()) {
s_logger.debug(vol + " of " + vm + " is READY, but template ids don't match, let the planner reassign a new pool");
continue;
}
final StoragePool pool = (StoragePool) dataStoreMgr.getPrimaryDataStore(vol.getPoolId());
if (!pool.isInMaintenance()) {
s_logger.debug("Root volume is ready, need to place VM in volume's cluster");
final long rootVolDcId = pool.getDataCenterId();
final Long rootVolPodId = pool.getPodId();
final Long rootVolClusterId = pool.getClusterId();
if (planToDeploy != null && planToDeploy.getDataCenterId() != 0) {
final Long clusterIdSpecified = planToDeploy.getClusterId();
if (clusterIdSpecified != null && rootVolClusterId != null) {
checkIfPlanIsDeployable(vm, rootVolClusterId, clusterIdSpecified);
}
plan = new DataCenterDeployment(planToDeploy.getDataCenterId(), planToDeploy.getPodId(), planToDeploy.getClusterId(), planToDeploy.getHostId(), vol.getPoolId(), null, ctx);
} else {
plan = new DataCenterDeployment(rootVolDcId, rootVolPodId, rootVolClusterId, null, vol.getPoolId(), null, ctx);
s_logger.debug(vol + " is READY, changing deployment plan to use this pool's dcId: " + rootVolDcId + " , podId: " + rootVolPodId + " , and clusterId: " + rootVolClusterId);
planChangedByVolume = true;
}
}
}
}
final Account owner = _entityMgr.findById(Account.class, vm.getAccountId());
final VirtualMachineProfileImpl vmProfile = new VirtualMachineProfileImpl(vm, template, offering, owner, params);
final DeployDestination dest = _dpMgr.planDeployment(vmProfile, plan, avoids, planner);
if (dest == null) {
if (planChangedByVolume) {
plan = originalPlan;
planChangedByVolume = false;
// do not enter volume reuse for next retry, since we want to look for resources outside the volume's cluster
reuseVolume = false;
continue;
}
throw new InsufficientServerCapacityException("Unable to create a deployment for " + vmProfile, DataCenter.class, plan.getDataCenterId(), areAffinityGroupsAssociated(vmProfile));
}
if (dest != null) {
avoids.addHost(dest.getHost().getId());
journal.record("Deployment found ", vmProfile, dest);
}
long destHostId = dest.getHost().getId();
vm.setPodIdToDeployIn(dest.getPod().getId());
final Long cluster_id = dest.getCluster().getId();
final ClusterDetailsVO cluster_detail_cpu = _clusterDetailsDao.findDetail(cluster_id, "cpuOvercommitRatio");
final ClusterDetailsVO cluster_detail_ram = _clusterDetailsDao.findDetail(cluster_id, "memoryOvercommitRatio");
// storing the value of overcommit in the vm_details table for doing a capacity check in case the cluster overcommit ratio is changed.
if (_uservmDetailsDao.findDetail(vm.getId(), "cpuOvercommitRatio") == null && (Float.parseFloat(cluster_detail_cpu.getValue()) > 1f || Float.parseFloat(cluster_detail_ram.getValue()) > 1f)) {
_uservmDetailsDao.addDetail(vm.getId(), "cpuOvercommitRatio", cluster_detail_cpu.getValue(), true);
_uservmDetailsDao.addDetail(vm.getId(), "memoryOvercommitRatio", cluster_detail_ram.getValue(), true);
} else if (_uservmDetailsDao.findDetail(vm.getId(), "cpuOvercommitRatio") != null) {
_uservmDetailsDao.addDetail(vm.getId(), "cpuOvercommitRatio", cluster_detail_cpu.getValue(), true);
_uservmDetailsDao.addDetail(vm.getId(), "memoryOvercommitRatio", cluster_detail_ram.getValue(), true);
}
vmProfile.setCpuOvercommitRatio(Float.parseFloat(cluster_detail_cpu.getValue()));
vmProfile.setMemoryOvercommitRatio(Float.parseFloat(cluster_detail_ram.getValue()));
StartAnswer startAnswer = null;
try {
if (!changeState(vm, Event.OperationRetry, destHostId, work, Step.Prepare)) {
throw new ConcurrentOperationException("Unable to update the state of the Virtual Machine " + vm.getUuid() + " oldstate: " + vm.getState() + "Event :" + Event.OperationRetry);
}
} catch (final NoTransitionException e1) {
throw new ConcurrentOperationException(e1.getMessage());
}
try {
s_logger.debug("VM is being created in podId: " + vm.getPodIdToDeployIn());
_networkMgr.prepare(vmProfile, dest, ctx);
volumeMgr.prepare(vmProfile, dest);
// since StorageMgr succeeded in volume creation, reuse Volume for further tries until current cluster has capacity
if (!reuseVolume) {
reuseVolume = true;
}
vmGuru.finalizeVirtualMachineProfile(vmProfile, dest, ctx);
final VirtualMachineTO vmTO = hvGuru.implement(vmProfile);
handlePath(vmTO.getDisks(), vm.getHypervisorType());
final Commands cmds = new Commands(Command.OnError.Stop);
cmds.addCommand(new StartCommand(vmTO, dest.getHost(), getExecuteInSequence(vm.getHypervisorType())));
vmGuru.finalizeDeployment(cmds, vmProfile, dest, ctx);
work = _workDao.findById(work.getId());
if (work == null || work.getStep() != Step.Prepare) {
throw new ConcurrentOperationException("Work steps have been changed: " + work);
}
_workDao.updateStep(work, Step.Starting);
_agentMgr.send(destHostId, cmds);
_workDao.updateStep(work, Step.Started);
startAnswer = cmds.getAnswer(StartAnswer.class);
// Set date and version we start this VM
vm.setLastStartDateTime(getCurrentLocalDateTimeStamp());
vm.setLastStartVersion(VirtualMachineManagerImpl.class.getPackage().getImplementationVersion());
// Reset VM compliance state
if (vm.getComplianceStatus() == ComplianceStatus.VMNeedsRestart) {
vm.setComplianceStatus(ComplianceStatus.Compliant);
}
_vmDao.persist(vm);
if (startAnswer != null && startAnswer.getResult()) {
handlePath(vmTO.getDisks(), startAnswer.getIqnToPath());
final String host_guid = startAnswer.getHost_guid();
if (host_guid != null) {
final HostVO finalHost = _resourceMgr.findHostByGuid(host_guid);
if (finalHost == null) {
throw new CloudRuntimeException("Host Guid " + host_guid + " doesn't exist in DB, something went wrong while processing start answer: " + startAnswer);
}
destHostId = finalHost.getId();
}
if (vmGuru.finalizeStart(vmProfile, destHostId, cmds, ctx)) {
syncDiskChainChange(startAnswer);
if (!changeState(vm, Event.OperationSucceeded, destHostId, work, Step.Done)) {
s_logger.error("Unable to transition to a new state. VM uuid: " + vm.getUuid() + "VM oldstate:" + vm.getState() + "Event:" + Event.OperationSucceeded);
throw new ConcurrentOperationException("Failed to deploy VM" + vm.getUuid());
}
// Update GPU device capacity
final GPUDeviceTO gpuDevice = startAnswer.getVirtualMachine().getGpuDevice();
if (gpuDevice != null) {
_resourceMgr.updateGPUDetails(destHostId, gpuDevice.getGroupDetails());
}
startedVm = vm;
s_logger.debug("Start completed for VM " + vm);
return;
} else {
s_logger.info("The guru did not like the answers so stopping " + vm);
final StopCommand cmd = new StopCommand(vm, getExecuteInSequence(vm.getHypervisorType()), false);
final Answer answer = _agentMgr.easySend(destHostId, cmd);
if (answer != null && answer instanceof StopAnswer) {
final StopAnswer stopAns = (StopAnswer) answer;
if (vm.getType() == VirtualMachineType.User) {
final String platform = stopAns.getPlatform();
if (platform != null) {
final Map<String, String> vmmetadata = new HashMap<>();
vmmetadata.put(vm.getInstanceName(), platform);
syncVMMetaData(vmmetadata);
}
}
}
if (answer == null || !answer.getResult()) {
s_logger.warn("Unable to stop " + vm + " due to " + (answer != null ? answer.getDetails() : "no answers"));
_haMgr.scheduleStop(vm, destHostId, HaWork.HaWorkType.ForceStop);
throw new ExecutionException("Unable to stop this VM, " + vm.getUuid() + " so we are unable to retry the start operation");
}
throw new ExecutionException("Unable to start VM:" + vm.getUuid() + " due to error in finalizeStart, not retrying");
}
}
s_logger.info("Unable to start VM on " + dest.getHost() + " due to " + (startAnswer == null ? " no start answer" : startAnswer.getDetails()));
if (startAnswer != null && startAnswer.getContextParam("stopRetry") != null) {
break;
}
} catch (final OperationTimedoutException e) {
s_logger.debug("Unable to send the start command to host " + dest.getHost() + " failed to start VM: " + vm.getUuid());
if (e.isActive()) {
_haMgr.scheduleStop(vm, destHostId, HaWork.HaWorkType.CheckStop);
}
canRetry = false;
throw new AgentUnavailableException("Unable to start " + vm.getHostName(), destHostId, e);
} catch (final ResourceUnavailableException e) {
s_logger.info("Unable to contact resource.", e);
if (!avoids.add(e)) {
if (e.getScope() == Volume.class || e.getScope() == Nic.class) {
throw e;
} else {
s_logger.warn("unexpected ResourceUnavailableException : " + e.getScope().getName(), e);
throw e;
}
}
} catch (final InsufficientCapacityException e) {
s_logger.info("Insufficient capacity ", e);
if (!avoids.add(e)) {
if (e.getScope() == Volume.class || e.getScope() == Nic.class) {
throw e;
} else {
s_logger.warn("unexpected InsufficientCapacityException : " + e.getScope().getName(), e);
}
}
} catch (final ExecutionException | NoTransitionException e) {
s_logger.error("Failed to start instance " + vm, e);
throw new AgentUnavailableException("Unable to start instance due to " + e.getMessage(), destHostId, e);
} finally {
if (startedVm == null && canRetry) {
final Step prevStep = work.getStep();
_workDao.updateStep(work, Step.Release);
// If previous step was started/ing && we got a valid answer
if ((prevStep == Step.Started || prevStep == Step.Starting) && startAnswer != null && startAnswer.getResult()) {
// TODO check the response of cleanup
// and record it in DB for retry
cleanup(vmGuru, vmProfile, work, Event.OperationFailed, false);
} else {
// if step is not starting/started, send cleanup command with force=true
cleanup(vmGuru, vmProfile, work, Event.OperationFailed, true);
}
}
}
}
} finally {
if (startedVm == null) {
if (canRetry) {
try {
changeState(vm, Event.OperationFailed, null, work, Step.Done);
} catch (final NoTransitionException e) {
throw new ConcurrentOperationException(e.getMessage());
}
}
}
if (planToDeploy != null) {
planToDeploy.setAvoids(avoids);
}
}
if (startedVm == null) {
throw new CloudRuntimeException("Unable to start instance '" + vm.getHostName() + "' (" + vm.getUuid() + "), see management server log for details");
}
}
use of com.cloud.legacymodel.exceptions.CloudRuntimeException in project cosmic by MissionCriticalCloud.
the class VolumeOrchestrator method copyVolumeFromSecToPrimary.
@DB
public VolumeInfo copyVolumeFromSecToPrimary(final VolumeInfo volume, final VirtualMachine vm, final VirtualMachineTemplate template, final DataCenter dc, final Pod pod, final Long clusterId, final DiskOffering diskOffering, final List<StoragePool> avoids) {
final HashSet<StoragePool> avoidPools = new HashSet<>(avoids);
final DiskProfile dskCh = createDiskCharacteristics(volume, template, dc, diskOffering);
dskCh.setHyperType(vm.getHypervisorType());
// Find a suitable storage to create volume on
final StoragePool destPool = findStoragePool(dskCh, dc, pod, clusterId, null, vm, avoidPools);
final DataStore destStore = this.dataStoreMgr.getDataStore(destPool.getId(), DataStoreRole.Primary);
final AsyncCallFuture<VolumeService.VolumeApiResult> future = this.volService.copyVolume(volume, destStore);
try {
final VolumeService.VolumeApiResult result = future.get();
if (result.isFailed()) {
s_logger.debug("copy volume failed: " + result.getResult());
throw new CloudRuntimeException("copy volume failed: " + result.getResult());
}
return result.getVolume();
} catch (final InterruptedException | ExecutionException e) {
s_logger.debug("Failed to copy volume: " + volume.getId(), e);
throw new CloudRuntimeException("Failed to copy volume", e);
}
}
use of com.cloud.legacymodel.exceptions.CloudRuntimeException in project cosmic by MissionCriticalCloud.
the class VolumeOrchestrator method getTasks.
private List<VolumeTask> getTasks(final List<VolumeVO> vols, final Map<Volume, StoragePool> destVols, final VirtualMachineProfile vm) throws StorageUnavailableException {
final boolean recreate = RecreatableSystemVmEnabled.value();
final List<VolumeTask> tasks = new ArrayList<>();
for (final VolumeVO vol : vols) {
StoragePoolVO assignedPool = null;
if (destVols != null) {
final StoragePool pool = destVols.get(vol);
if (pool != null) {
assignedPool = this._storagePoolDao.findById(pool.getId());
}
}
if (assignedPool == null && recreate) {
assignedPool = this._storagePoolDao.findById(vol.getPoolId());
}
if (assignedPool != null) {
final Volume.State state = vol.getState();
if (state == Volume.State.Allocated || state == Volume.State.Creating) {
final VolumeTask task = new VolumeTask(VolumeTaskType.RECREATE, vol, null);
tasks.add(task);
} else {
if (vol.isRecreatable()) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("Volume " + vol + " will be recreated on storage pool " + assignedPool + " assigned by deploymentPlanner");
}
final VolumeTask task = new VolumeTask(VolumeTaskType.RECREATE, vol, null);
tasks.add(task);
} else {
if (assignedPool.getId() != vol.getPoolId()) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("Mismatch in storage pool " + assignedPool + " assigned by deploymentPlanner and the one associated with volume " + vol);
}
final DiskOffering diskOffering = this._entityMgr.findById(DiskOffering.class, vol.getDiskOfferingId());
if (diskOffering.getUseLocalStorage()) {
// Currently migration of local volume is not supported so bail out
if (s_logger.isDebugEnabled()) {
s_logger.debug("Local volume " + vol + " cannot be recreated on storagepool " + assignedPool + " assigned by deploymentPlanner");
}
throw new CloudRuntimeException("Local volume " + vol + " cannot be recreated on storagepool " + assignedPool + " assigned by deploymentPlanner");
} else {
// Check if storage migration is enabled in config
final Boolean isHAOperation = (Boolean) vm.getParameter(VirtualMachineProfile.Param.HaOperation);
Boolean storageMigrationEnabled = true;
if (isHAOperation != null && isHAOperation) {
storageMigrationEnabled = StorageHAMigrationEnabled.value();
} else {
storageMigrationEnabled = StorageMigrationEnabled.value();
}
// Always allow ISOs volumes to be "migrated"
if (storageMigrationEnabled || vol.getIsoId() != null) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("Shared volume " + vol + " will be migrated on storage pool " + assignedPool + " assigned by deploymentPlanner");
}
final VolumeTask task = new VolumeTask(VolumeTaskType.MIGRATE, vol, assignedPool);
tasks.add(task);
} else {
throw new CloudRuntimeException("Cannot start VM on the hypervisor it was last running on, due to not enough capacity. Please try to start on" + " " + "another hypervisor in the same cluster, or migrate the volumes to another storage pool. Automatic Volume Migration is disabled, " + "so this is not handled automatically.");
}
}
} else {
final StoragePoolVO pool = this._storagePoolDao.findById(vol.getPoolId());
final VolumeTask task = new VolumeTask(VolumeTaskType.NOP, vol, pool);
tasks.add(task);
}
}
}
} else {
if (vol.getPoolId() == null) {
throw new StorageUnavailableException("Volume has no pool associate and also no storage pool assigned in DeployDestination, Unable to create " + vol, Volume.class, vol.getId());
}
if (s_logger.isDebugEnabled()) {
s_logger.debug("No need to recreate the volume: " + vol + ", since it already has a pool assigned: " + vol.getPoolId() + ", adding disk to VM");
}
final StoragePoolVO pool = this._storagePoolDao.findById(vol.getPoolId());
final VolumeTask task = new VolumeTask(VolumeTaskType.NOP, vol, pool);
tasks.add(task);
}
}
return tasks;
}
use of com.cloud.legacymodel.exceptions.CloudRuntimeException in project cosmic by MissionCriticalCloud.
the class VolumeOrchestrator method destroyVolume.
@Override
public void destroyVolume(final Volume volume) {
try {
// Mark volume as removed if volume has not been created on primary
if (volume.getState() == Volume.State.Allocated) {
this._volsDao.remove(volume.getId());
stateTransitTo(volume, Volume.Event.DestroyRequested);
} else {
this.volService.destroyVolume(volume.getId());
}
this._resourceLimitMgr.decrementResourceCount(volume.getAccountId(), ResourceType.volume, volume.isDisplay());
// FIXME - why recalculate and not decrement
this._resourceLimitMgr.recalculateResourceCount(volume.getAccountId(), volume.getDomainId(), ResourceType.primary_storage.getOrdinal());
} catch (final Exception e) {
s_logger.debug("Failed to destroy volume" + volume.getId(), e);
throw new CloudRuntimeException("Failed to destroy volume" + volume.getId(), e);
}
}
Aggregations