use of com.cloud.exception.InsufficientServerCapacityException in project cloudstack by apache.
the class VirtualMachineManagerImpl method migrateAway.
@Override
public void migrateAway(final String vmUuid, final long srcHostId) throws InsufficientServerCapacityException {
final AsyncJobExecutionContext jobContext = AsyncJobExecutionContext.getCurrentExecutionContext();
if (jobContext.isJobDispatchedBy(VmWorkConstants.VM_WORK_JOB_DISPATCHER)) {
final VirtualMachine vm = _vmDao.findByUuid(vmUuid);
VmWorkJobVO placeHolder = createPlaceHolderWork(vm.getId());
try {
try {
orchestrateMigrateAway(vmUuid, srcHostId, null);
} catch (final InsufficientServerCapacityException e) {
s_logger.warn("Failed to deploy vm " + vmUuid + " with original planner, sending HAPlanner");
orchestrateMigrateAway(vmUuid, srcHostId, _haMgr.getHAPlanner());
}
} finally {
_workJobDao.expunge(placeHolder.getId());
}
} else {
final Outcome<VirtualMachine> outcome = migrateVmAwayThroughJobQueue(vmUuid, srcHostId);
retrieveVmFromJobOutcome(outcome, vmUuid, "migrateVmAway");
try {
retrieveResultFromJobOutcomeAndThrowExceptionIfNeeded(outcome);
} catch (ResourceUnavailableException | InsufficientCapacityException ex) {
throw new RuntimeException("Unexpected exception", ex);
}
}
}
use of com.cloud.exception.InsufficientServerCapacityException in project cloudstack by apache.
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 VirtualMachineGuru vmGuru = getVmGuru(vm);
final Ternary<VMInstanceVO, ReservationContext, ItWorkVO> start = changeToStartState(vmGuru, 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());
DataCenterDeployment plan = new DataCenterDeployment(vm.getDataCenterId(), vm.getPodIdToDeployIn(), null, null, null, null, ctx);
if (planToDeploy != null && planToDeploy.getDataCenterId() != 0) {
if (s_logger.isDebugEnabled()) {
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());
// check resource count if ResourceCountRunningVMsonly.value() = true
final Account owner = _entityMgr.findById(Account.class, vm.getAccountId());
if (VirtualMachine.Type.User.equals(vm.type) && ResourceCountRunningVMsonly.value()) {
resourceCountIncrement(owner.getAccountId(), new Long(offering.getCpu()), new Long(offering.getRamSize()));
}
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();
}
if (s_logger.isDebugEnabled()) {
s_logger.debug("Deploy avoids pods: " + avoids.getPodsToAvoid() + ", clusters: " + avoids.getClustersToAvoid() + ", hosts: " + avoids.getHostsToAvoid());
}
boolean planChangedByVolume = false;
boolean reuseVolume = true;
final DataCenterDeployment originalPlan = plan;
int retry = StartRetry.value();
while (retry-- != 0) {
s_logger.debug("VM start attempt #" + (StartRetry.value() - retry));
if (reuseVolume) {
final List<VolumeVO> vols = _volsDao.findReadyRootVolumesByInstance(vm.getId());
for (final VolumeVO vol : vols) {
final Long volTemplateId = vol.getTemplateId();
if (volTemplateId != null && volTemplateId != template.getId()) {
if (s_logger.isDebugEnabled()) {
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()) {
if (s_logger.isDebugEnabled()) {
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) {
if (!rootVolClusterId.equals(clusterIdSpecified)) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("Cannot satisfy the deployment plan passed in since the ready Root volume is in different cluster. volume's cluster: " + rootVolClusterId + ", cluster specified: " + clusterIdSpecified);
}
throw new ResourceUnavailableException("Root volume is ready in different cluster, Deployment plan provided cannot be satisfied, unable to create a deployment for " + vm, Cluster.class, 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);
if (s_logger.isDebugEnabled()) {
s_logger.debug(vol + " is READY, changing deployment plan to use this pool's dcId: " + rootVolDcId + " , podId: " + rootVolPodId + " , and clusterId: " + rootVolClusterId);
}
planChangedByVolume = true;
}
}
}
}
final VirtualMachineProfileImpl vmProfile = new VirtualMachineProfileImpl(vm, template, offering, owner, params);
logBootModeParameters(params);
DeployDestination dest = null;
try {
dest = _dpMgr.planDeployment(vmProfile, plan, avoids, planner);
} catch (final AffinityConflictException e2) {
s_logger.warn("Unable to create deployment, affinity rules associted to the VM conflict", e2);
throw new CloudRuntimeException("Unable to create deployment, affinity rules associted to the VM conflict");
}
if (dest == null) {
if (planChangedByVolume) {
plan = originalPlan;
planChangedByVolume = false;
reuseVolume = false;
continue;
}
throw new InsufficientServerCapacityException("Unable to create a deployment for " + vmProfile, DataCenter.class, plan.getDataCenterId(), areAffinityGroupsAssociated(vmProfile));
}
avoids.addHost(dest.getHost().getId());
if (!template.isDeployAsIs()) {
journal.record("Deployment found - Attempt #" + (StartRetry.value() - retry), 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, VmDetailConstants.CPU_OVER_COMMIT_RATIO);
final ClusterDetailsVO cluster_detail_ram = _clusterDetailsDao.findDetail(cluster_id, VmDetailConstants.MEMORY_OVER_COMMIT_RATIO);
if (userVmDetailsDao.findDetail(vm.getId(), VmDetailConstants.CPU_OVER_COMMIT_RATIO) == null && (Float.parseFloat(cluster_detail_cpu.getValue()) > 1f || Float.parseFloat(cluster_detail_ram.getValue()) > 1f)) {
userVmDetailsDao.addDetail(vm.getId(), VmDetailConstants.CPU_OVER_COMMIT_RATIO, cluster_detail_cpu.getValue(), true);
userVmDetailsDao.addDetail(vm.getId(), VmDetailConstants.MEMORY_OVER_COMMIT_RATIO, cluster_detail_ram.getValue(), true);
} else if (userVmDetailsDao.findDetail(vm.getId(), VmDetailConstants.CPU_OVER_COMMIT_RATIO) != null) {
userVmDetailsDao.addDetail(vm.getId(), VmDetailConstants.CPU_OVER_COMMIT_RATIO, cluster_detail_cpu.getValue(), true);
userVmDetailsDao.addDetail(vm.getId(), VmDetailConstants.MEMORY_OVER_COMMIT_RATIO, 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 {
resetVmNicsDeviceId(vm.getId());
_networkMgr.prepare(vmProfile, dest, ctx);
if (vm.getHypervisorType() != HypervisorType.BareMetal) {
volumeMgr.prepare(vmProfile, dest);
}
if (!reuseVolume) {
reuseVolume = true;
}
vmGuru.finalizeVirtualMachineProfile(vmProfile, dest, ctx);
final VirtualMachineTO vmTO = hvGuru.implement(vmProfile);
checkAndSetEnterSetupMode(vmTO, params);
handlePath(vmTO.getDisks(), vm.getHypervisorType());
Commands cmds = new Commands(Command.OnError.Stop);
cmds.addCommand(new StartCommand(vmTO, dest.getHost(), getExecuteInSequence(vm.getHypervisorType())));
vmGuru.finalizeDeployment(cmds, vmProfile, dest, ctx);
addExtraConfig(vmTO);
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);
if (startAnswer != null && startAnswer.getResult()) {
handlePath(vmTO.getDisks(), startAnswer.getIqnToData());
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());
}
final GPUDeviceTO gpuDevice = startAnswer.getVirtualMachine().getGpuDevice();
if (gpuDevice != null) {
_resourceMgr.updateGPUDetails(destHostId, gpuDevice.getGroupDetails());
}
if (userVmDetailsDao.findDetail(vm.getId(), VmDetailConstants.DEPLOY_VM) != null) {
userVmDetailsDao.removeDetail(vm.getId(), VmDetailConstants.DEPLOY_VM);
}
startedVm = vm;
if (s_logger.isDebugEnabled()) {
s_logger.debug("Start completed for VM " + vm);
}
final Host vmHost = _hostDao.findById(destHostId);
if (vmHost != null && (VirtualMachine.Type.ConsoleProxy.equals(vm.getType()) || VirtualMachine.Type.SecondaryStorageVm.equals(vm.getType())) && caManager.canProvisionCertificates()) {
final Map<String, String> sshAccessDetails = _networkMgr.getSystemVMAccessDetails(vm);
for (int retries = 3; retries > 0; retries--) {
try {
setupAgentSecurity(vmHost, sshAccessDetails, vm);
return;
} catch (final AgentUnavailableException | OperationTimedoutException e) {
s_logger.error("Retrying after catching exception while trying to secure agent for systemvm id=" + vm.getId(), e);
}
}
throw new CloudRuntimeException("Failed to setup and secure agent for systemvm id=" + vm.getId());
}
return;
} else {
if (s_logger.isDebugEnabled()) {
s_logger.info("The guru did not like the answers so stopping " + vm);
}
StopCommand stopCmd = new StopCommand(vm, getExecuteInSequence(vm.getHypervisorType()), false);
stopCmd.setControlIp(getControlNicIpForVM(vm));
Map<String, Boolean> vlanToPersistenceMap = getVlanToPersistenceMapForVM(vm.getId());
if (MapUtils.isNotEmpty(vlanToPersistenceMap)) {
stopCmd.setVlanToPersistenceMap(vlanToPersistenceMap);
}
final StopCommand cmd = stopCmd;
final Answer answer = _agentMgr.easySend(destHostId, cmd);
if (answer != null && answer instanceof StopAnswer) {
final StopAnswer stopAns = (StopAnswer) answer;
if (vm.getType() == VirtualMachine.Type.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, WorkType.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 (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, WorkType.CheckStop);
}
canRetry = false;
throw new AgentUnavailableException("Unable to start " + vm.getHostName(), destHostId, e);
} catch (final ResourceUnavailableException e) {
s_logger.warn("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.warn("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 (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);
} catch (final StorageAccessException e) {
s_logger.warn("Unable to access storage on host", e);
} finally {
if (startedVm == null && canRetry) {
final Step prevStep = work.getStep();
_workDao.updateStep(work, Step.Release);
if ((prevStep == Step.Started || prevStep == Step.Starting) && startAnswer != null && startAnswer.getResult()) {
cleanup(vmGuru, vmProfile, work, Event.OperationFailed, false);
} else {
cleanup(vmGuru, vmProfile, work, Event.OperationFailed, true);
}
}
}
}
} finally {
if (startedVm == null) {
if (VirtualMachine.Type.User.equals(vm.type) && ResourceCountRunningVMsonly.value()) {
resourceCountDecrement(owner.getAccountId(), new Long(offering.getCpu()), new Long(offering.getRamSize()));
}
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.exception.InsufficientServerCapacityException in project cloudstack by apache.
the class HighAvailabilityManagerImpl method migrate.
public Long migrate(final HaWorkVO work) {
long vmId = work.getInstanceId();
long srcHostId = work.getHostId();
VMInstanceVO vm = _instanceDao.findById(vmId);
if (vm == null) {
s_logger.info("Unable to find vm: " + vmId + ", skipping migrate.");
return null;
}
s_logger.info("Migration attempt: for VM " + vm.getUuid() + "from host id " + srcHostId + ". Starting attempt: " + (1 + work.getTimesTried()) + "/" + _maxRetries + " times.");
try {
work.setStep(Step.Migrating);
_haDao.update(work.getId(), work);
// First try starting the vm with its original planner, if it doesn't succeed send HAPlanner as its an emergency.
_itMgr.migrateAway(vm.getUuid(), srcHostId);
return null;
} catch (InsufficientServerCapacityException e) {
s_logger.warn("Migration attempt: Insufficient capacity for migrating a VM " + vm.getUuid() + " from source host id " + srcHostId + ". Exception: " + e.getMessage());
_resourceMgr.migrateAwayFailed(srcHostId, vmId);
return (System.currentTimeMillis() >> 10) + _migrateRetryInterval;
} catch (Exception e) {
s_logger.warn("Migration attempt: Unexpected exception occurred when attempting migration of " + vm.getUuid() + e.getMessage());
throw e;
}
}
use of com.cloud.exception.InsufficientServerCapacityException in project cloudstack by apache.
the class KubernetesClusterManagerImpl method plan.
private DeployDestination plan(final long nodesCount, final DataCenter zone, final ServiceOffering offering) throws InsufficientServerCapacityException {
final int cpu_requested = offering.getCpu() * offering.getSpeed();
final long ram_requested = offering.getRamSize() * 1024L * 1024L;
List<HostVO> hosts = resourceManager.listAllHostsInOneZoneByType(Type.Routing, zone.getId());
final Map<String, Pair<HostVO, Integer>> hosts_with_resevered_capacity = new ConcurrentHashMap<String, Pair<HostVO, Integer>>();
for (HostVO h : hosts) {
hosts_with_resevered_capacity.put(h.getUuid(), new Pair<HostVO, Integer>(h, 0));
}
boolean suitable_host_found = false;
Cluster planCluster = null;
for (int i = 1; i <= nodesCount; i++) {
suitable_host_found = false;
for (Map.Entry<String, Pair<HostVO, Integer>> hostEntry : hosts_with_resevered_capacity.entrySet()) {
Pair<HostVO, Integer> hp = hostEntry.getValue();
HostVO h = hp.first();
hostDao.loadHostTags(h);
if (StringUtils.isNotEmpty(offering.getHostTag()) && !(h.getHostTags() != null && h.getHostTags().contains(offering.getHostTag()))) {
continue;
}
int reserved = hp.second();
reserved++;
ClusterVO cluster = clusterDao.findById(h.getClusterId());
ClusterDetailsVO cluster_detail_cpu = clusterDetailsDao.findDetail(cluster.getId(), "cpuOvercommitRatio");
ClusterDetailsVO cluster_detail_ram = clusterDetailsDao.findDetail(cluster.getId(), "memoryOvercommitRatio");
Float cpuOvercommitRatio = Float.parseFloat(cluster_detail_cpu.getValue());
Float memoryOvercommitRatio = Float.parseFloat(cluster_detail_ram.getValue());
if (LOGGER.isDebugEnabled()) {
LOGGER.debug(String.format("Checking host ID: %s for capacity already reserved %d", h.getUuid(), reserved));
}
if (capacityManager.checkIfHostHasCapacity(h.getId(), cpu_requested * reserved, ram_requested * reserved, false, cpuOvercommitRatio, memoryOvercommitRatio, true)) {
if (LOGGER.isDebugEnabled()) {
LOGGER.debug(String.format("Found host ID: %s for with enough capacity, CPU=%d RAM=%s", h.getUuid(), cpu_requested * reserved, toHumanReadableSize(ram_requested * reserved)));
}
hostEntry.setValue(new Pair<HostVO, Integer>(h, reserved));
suitable_host_found = true;
planCluster = cluster;
break;
}
}
if (!suitable_host_found) {
if (LOGGER.isInfoEnabled()) {
LOGGER.info(String.format("Suitable hosts not found in datacenter ID: %s for node %d with offering ID: %s", zone.getUuid(), i, offering.getUuid()));
}
break;
}
}
if (suitable_host_found) {
if (LOGGER.isInfoEnabled()) {
LOGGER.info(String.format("Suitable hosts found in datacenter ID: %s, creating deployment destination", zone.getUuid()));
}
return new DeployDestination(zone, null, planCluster, null);
}
String msg = String.format("Cannot find enough capacity for Kubernetes cluster(requested cpu=%d memory=%s) with offering ID: %s", cpu_requested * nodesCount, toHumanReadableSize(ram_requested * nodesCount), offering.getUuid());
LOGGER.warn(msg);
throw new InsufficientServerCapacityException(msg, DataCenter.class, zone.getId());
}
use of com.cloud.exception.InsufficientServerCapacityException in project cloudstack by apache.
the class ResourceManagerImpl method migrateAwayVmWithVolumes.
/**
* Looks for Hosts able to allocate the VM and migrates the VM with its volume.
*/
private void migrateAwayVmWithVolumes(HostVO host, VMInstanceVO vm) {
final DataCenterDeployment plan = new DataCenterDeployment(host.getDataCenterId(), host.getPodId(), host.getClusterId(), null, null, null);
ServiceOfferingVO offeringVO = serviceOfferingDao.findById(vm.getServiceOfferingId());
final VirtualMachineProfile profile = new VirtualMachineProfileImpl(vm, null, offeringVO, null, null);
plan.setMigrationPlan(true);
DeployDestination dest = null;
try {
dest = deploymentManager.planDeployment(profile, plan, new DeploymentPlanner.ExcludeList(), null);
} catch (InsufficientServerCapacityException e) {
throw new CloudRuntimeException(String.format("Maintenance failed, could not find deployment destination for VM [id=%s, name=%s].", vm.getId(), vm.getInstanceName()), e);
}
Host destHost = dest.getHost();
try {
_vmMgr.migrateWithStorage(vm.getUuid(), host.getId(), destHost.getId(), null);
} catch (ResourceUnavailableException e) {
throw new CloudRuntimeException(String.format("Maintenance failed, could not migrate VM [id=%s, name=%s] with local storage from host [id=%s, name=%s] to host [id=%s, name=%s].", vm.getId(), vm.getInstanceName(), host.getId(), host.getName(), destHost.getId(), destHost.getName()), e);
}
}
Aggregations