@Override
public TemplateProfile prepare(RegisterTemplateCmd cmd) throws ResourceAllocationException {
TemplateProfile profile = super.prepare(cmd);
if (profile.getZoneId() == null || profile.getZoneId() == -1) {
List<DataCenterVO> dcs = _dcDao.listAllIncludingRemoved();
for (DataCenterVO dc : dcs) {
List<HostVO> pxeServers =
_resourceMgr.listAllHostsInOneZoneByType(Host.Type.BaremetalPxe, dc.getId());
if (pxeServers.size() == 0) {
throw new CloudRuntimeException(
"Please add PXE server before adding baremetal template in zone " + dc.getName());
}
}
} else {
List<HostVO> pxeServers =
_resourceMgr.listAllHostsInOneZoneByType(Host.Type.BaremetalPxe, profile.getZoneId());
if (pxeServers.size() == 0) {
throw new CloudRuntimeException(
"Please add PXE server before adding baremetal template in zone "
+ profile.getZoneId());
}
}
return profile;
}
@Override
public void scheduleRestartForVmsOnHost(final HostVO host) {
if (host.getType() != Host.Type.Routing) {
return;
}
s_logger.warn("Scheduling restart for VMs on host " + host.getId());
final List<VMInstanceVO> vms = _instanceDao.listByHostId(host.getId());
final DataCenterVO dcVO = _dcDao.findById(host.getDataCenterId());
// send an email alert that the host is down
StringBuilder sb = null;
if ((vms != null) && !vms.isEmpty()) {
sb = new StringBuilder();
sb.append(" Starting HA on the following VMs: ");
// collect list of vm names for the alert email
VMInstanceVO vm = vms.get(0);
if (vm.isHaEnabled()) {
sb.append(" " + vm.getName());
}
for (int i = 1; i < vms.size(); i++) {
vm = vms.get(i);
if (vm.isHaEnabled()) {
sb.append(" " + vm.getName());
}
}
}
// send an email alert that the host is down, include VMs
HostPodVO podVO = _podDao.findById(host.getPodId());
String hostDesc =
"name: "
+ host.getName()
+ " (id:"
+ host.getId()
+ "), availability zone: "
+ dcVO.getName()
+ ", pod: "
+ podVO.getName();
_alertMgr.sendAlert(
AlertManager.ALERT_TYPE_HOST,
host.getDataCenterId(),
host.getPodId(),
"Host is down, " + hostDesc,
"Host [" + hostDesc + "] is down." + ((sb != null) ? sb.toString() : ""));
for (final VMInstanceVO vm : vms) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("Notifying HA Mgr of to investigate vm " + vm.getId() + "-" + vm.getName());
}
scheduleRestart(vm, true);
}
}
@Override
@DB
public boolean delete(TemplateProfile profile) {
boolean success = true;
VMTemplateVO template = profile.getTemplate();
Long zoneId = profile.getZoneId();
Long templateId = template.getId();
String zoneName;
List<HostVO> secondaryStorageHosts;
if (!template.isCrossZones() && zoneId != null) {
DataCenterVO zone = _dcDao.findById(zoneId);
zoneName = zone.getName();
secondaryStorageHosts = _ssvmMgr.listSecondaryStorageHostsInOneZone(zoneId);
} else {
zoneName = "(all zones)";
secondaryStorageHosts = _ssvmMgr.listSecondaryStorageHostsInAllZones();
}
s_logger.debug(
"Attempting to mark template host refs for template: "
+ template.getName()
+ " as destroyed in zone: "
+ zoneName);
// Make sure the template is downloaded to all the necessary secondary storage hosts
for (HostVO secondaryStorageHost : secondaryStorageHosts) {
long hostId = secondaryStorageHost.getId();
List<VMTemplateHostVO> templateHostVOs = _tmpltHostDao.listByHostTemplate(hostId, templateId);
for (VMTemplateHostVO templateHostVO : templateHostVOs) {
if (templateHostVO.getDownloadState() == Status.DOWNLOAD_IN_PROGRESS) {
String errorMsg = "Please specify a template that is not currently being downloaded.";
s_logger.debug(
"Template: "
+ template.getName()
+ " is currently being downloaded to secondary storage host: "
+ secondaryStorageHost.getName()
+ "; cant' delete it.");
throw new CloudRuntimeException(errorMsg);
}
}
}
Account account = _accountDao.findByIdIncludingRemoved(template.getAccountId());
String eventType = "";
if (template.getFormat().equals(ImageFormat.ISO)) {
eventType = EventTypes.EVENT_ISO_DELETE;
} else {
eventType = EventTypes.EVENT_TEMPLATE_DELETE;
}
// Iterate through all necessary secondary storage hosts and mark the template on each host as
// destroyed
for (HostVO secondaryStorageHost : secondaryStorageHosts) {
long hostId = secondaryStorageHost.getId();
long sZoneId = secondaryStorageHost.getDataCenterId();
List<VMTemplateHostVO> templateHostVOs = _tmpltHostDao.listByHostTemplate(hostId, templateId);
for (VMTemplateHostVO templateHostVO : templateHostVOs) {
VMTemplateHostVO lock = _tmpltHostDao.acquireInLockTable(templateHostVO.getId());
try {
if (lock == null) {
s_logger.debug(
"Failed to acquire lock when deleting templateHostVO with ID: "
+ templateHostVO.getId());
success = false;
break;
}
UsageEventVO usageEvent =
new UsageEventVO(eventType, account.getId(), sZoneId, templateId, null);
_usageEventDao.persist(usageEvent);
templateHostVO.setDestroyed(true);
_tmpltHostDao.update(templateHostVO.getId(), templateHostVO);
String installPath = templateHostVO.getInstallPath();
if (installPath != null) {
Answer answer =
_agentMgr.sendToSecStorage(
secondaryStorageHost,
new DeleteTemplateCommand(secondaryStorageHost.getStorageUrl(), installPath));
if (answer == null || !answer.getResult()) {
s_logger.debug(
"Failed to delete "
+ templateHostVO
+ " due to "
+ ((answer == null) ? "answer is null" : answer.getDetails()));
} else {
_tmpltHostDao.remove(templateHostVO.getId());
s_logger.debug("Deleted template at: " + installPath);
}
} else {
_tmpltHostDao.remove(templateHostVO.getId());
}
VMTemplateZoneVO templateZone = _tmpltZoneDao.findByZoneTemplate(sZoneId, templateId);
if (templateZone != null) {
_tmpltZoneDao.remove(templateZone.getId());
}
} finally {
if (lock != null) {
_tmpltHostDao.releaseFromLockTable(lock.getId());
}
}
}
if (!success) {
break;
}
}
s_logger.debug(
"Successfully marked template host refs for template: "
+ template.getName()
+ " as destroyed in zone: "
+ zoneName);
// If there are no more non-destroyed template host entries for this template, delete it
if (success && (_tmpltHostDao.listByTemplateId(templateId).size() == 0)) {
long accountId = template.getAccountId();
VMTemplateVO lock = _tmpltDao.acquireInLockTable(templateId);
try {
if (lock == null) {
s_logger.debug("Failed to acquire lock when deleting template with ID: " + templateId);
success = false;
} else if (_tmpltDao.remove(templateId)) {
// Decrement the number of templates
_resourceLimitMgr.decrementResourceCount(accountId, ResourceType.template);
}
} finally {
if (lock != null) {
_tmpltDao.releaseFromLockTable(lock.getId());
}
}
s_logger.debug(
"Removed template: "
+ template.getName()
+ " because all of its template host refs were marked as destroyed.");
}
return success;
}
protected Long restart(HaWorkVO work) {
List<HaWorkVO> items = _haDao.listFutureHaWorkForVm(work.getInstanceId(), work.getId());
if (items.size() > 0) {
StringBuilder str =
new StringBuilder(
"Cancelling this work item because newer ones have been scheduled. Work Ids = [");
for (HaWorkVO item : items) {
str.append(item.getId()).append(", ");
}
str.delete(str.length() - 2, str.length()).append("]");
s_logger.info(str.toString());
return null;
}
items = _haDao.listRunningHaWorkForVm(work.getInstanceId());
if (items.size() > 0) {
StringBuilder str =
new StringBuilder(
"Waiting because there's HA work being executed on an item currently. Work Ids =[");
for (HaWorkVO item : items) {
str.append(item.getId()).append(", ");
}
str.delete(str.length() - 2, str.length()).append("]");
s_logger.info(str.toString());
return (System.currentTimeMillis() >> 10) + _investigateRetryInterval;
}
long vmId = work.getInstanceId();
VMInstanceVO vm = _itMgr.findByIdAndType(work.getType(), work.getInstanceId());
if (vm == null) {
s_logger.info("Unable to find vm: " + vmId);
return null;
}
s_logger.info("HA on " + vm);
if (vm.getState() != work.getPreviousState() || vm.getUpdated() != work.getUpdateTime()) {
s_logger.info(
"VM "
+ vm
+ " has been changed. Current State = "
+ vm.getState()
+ " Previous State = "
+ work.getPreviousState()
+ " last updated = "
+ vm.getUpdated()
+ " previous updated = "
+ work.getUpdateTime());
return null;
}
short alertType = AlertManager.ALERT_TYPE_USERVM;
if (VirtualMachine.Type.DomainRouter.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_DOMAIN_ROUTER;
} else if (VirtualMachine.Type.ConsoleProxy.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_CONSOLE_PROXY;
} else if (VirtualMachine.Type.SecondaryStorageVm.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_SSVM;
}
HostVO host = _hostDao.findById(work.getHostId());
boolean isHostRemoved = false;
if (host == null) {
host = _hostDao.findByIdIncludingRemoved(work.getHostId());
if (host != null) {
s_logger.debug(
"VM "
+ vm.toString()
+ " is now no longer on host "
+ work.getHostId()
+ " as the host is removed");
isHostRemoved = true;
}
}
DataCenterVO dcVO = _dcDao.findById(host.getDataCenterId());
HostPodVO podVO = _podDao.findById(host.getPodId());
String hostDesc =
"name: "
+ host.getName()
+ "(id:"
+ host.getId()
+ "), availability zone: "
+ dcVO.getName()
+ ", pod: "
+ podVO.getName();
Boolean alive = null;
if (work.getStep() == Step.Investigating) {
if (!isHostRemoved) {
if (vm.getHostId() == null || vm.getHostId() != work.getHostId()) {
s_logger.info("VM " + vm.toString() + " is now no longer on host " + work.getHostId());
return null;
}
Enumeration<Investigator> en = _investigators.enumeration();
Investigator investigator = null;
while (en.hasMoreElements()) {
investigator = en.nextElement();
alive = investigator.isVmAlive(vm, host);
s_logger.info(investigator.getName() + " found " + vm + "to be alive? " + alive);
if (alive != null) {
break;
}
}
boolean fenced = false;
if (alive == null) {
s_logger.debug("Fencing off VM that we don't know the state of");
Enumeration<FenceBuilder> enfb = _fenceBuilders.enumeration();
while (enfb.hasMoreElements()) {
FenceBuilder fb = enfb.nextElement();
Boolean result = fb.fenceOff(vm, host);
s_logger.info("Fencer " + fb.getName() + " returned " + result);
if (result != null && result) {
fenced = true;
break;
}
}
} else if (!alive) {
fenced = true;
} else {
s_logger.debug(
"VM " + vm.getHostName() + " is found to be alive by " + investigator.getName());
if (host.getStatus() == Status.Up) {
s_logger.info(vm + " is alive and host is up. No need to restart it.");
return null;
} else {
s_logger.debug("Rescheduling because the host is not up but the vm is alive");
return (System.currentTimeMillis() >> 10) + _investigateRetryInterval;
}
}
if (!fenced) {
s_logger.debug("We were unable to fence off the VM " + vm);
_alertMgr.sendAlert(
alertType,
vm.getDataCenterIdToDeployIn(),
vm.getPodIdToDeployIn(),
"Unable to restart " + vm.getHostName() + " which was running on host " + hostDesc,
"Insufficient capacity to restart VM, name: "
+ vm.getHostName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
return (System.currentTimeMillis() >> 10) + _restartRetryInterval;
}
try {
_itMgr.advanceStop(vm, true, _accountMgr.getSystemUser(), _accountMgr.getSystemAccount());
} catch (ResourceUnavailableException e) {
assert false : "How do we hit this when force is true?";
throw new CloudRuntimeException("Caught exception even though it should be handled.", e);
} catch (OperationTimedoutException e) {
assert false : "How do we hit this when force is true?";
throw new CloudRuntimeException("Caught exception even though it should be handled.", e);
} catch (ConcurrentOperationException e) {
assert false : "How do we hit this when force is true?";
throw new CloudRuntimeException("Caught exception even though it should be handled.", e);
}
work.setStep(Step.Scheduled);
_haDao.update(work.getId(), work);
} else {
s_logger.debug(
"How come that HA step is Investigating and the host is removed? Calling forced Stop on Vm anyways");
try {
_itMgr.advanceStop(vm, true, _accountMgr.getSystemUser(), _accountMgr.getSystemAccount());
} catch (ResourceUnavailableException e) {
assert false : "How do we hit this when force is true?";
throw new CloudRuntimeException("Caught exception even though it should be handled.", e);
} catch (OperationTimedoutException e) {
assert false : "How do we hit this when force is true?";
throw new CloudRuntimeException("Caught exception even though it should be handled.", e);
} catch (ConcurrentOperationException e) {
assert false : "How do we hit this when force is true?";
throw new CloudRuntimeException("Caught exception even though it should be handled.", e);
}
}
}
vm = _itMgr.findByIdAndType(vm.getType(), vm.getId());
if (!_forceHA && !vm.isHaEnabled()) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("VM is not HA enabled so we're done.");
}
return null; // VM doesn't require HA
}
if (!_storageMgr.canVmRestartOnAnotherServer(vm.getId())) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("VM can not restart on another server.");
}
return null;
}
if (work.getTimesTried() > _maxRetries) {
s_logger.warn("Retried to max times so deleting: " + vmId);
return null;
}
try {
VMInstanceVO started =
_itMgr.advanceStart(
vm,
new HashMap<VirtualMachineProfile.Param, Object>(),
_accountMgr.getSystemUser(),
_accountMgr.getSystemAccount());
if (started != null) {
s_logger.info("VM is now restarted: " + vmId + " on " + started.getHostId());
return null;
}
if (s_logger.isDebugEnabled()) {
s_logger.debug(
"Rescheduling VM " + vm.toString() + " to try again in " + _restartRetryInterval);
}
} catch (final InsufficientCapacityException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterIdToDeployIn(),
vm.getPodIdToDeployIn(),
"Unable to restart " + vm.getHostName() + " which was running on host " + hostDesc,
"Insufficient capacity to restart VM, name: "
+ vm.getHostName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
} catch (final ResourceUnavailableException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterIdToDeployIn(),
vm.getPodIdToDeployIn(),
"Unable to restart " + vm.getHostName() + " which was running on host " + hostDesc,
"The Storage is unavailable for trying to restart VM, name: "
+ vm.getHostName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
} catch (ConcurrentOperationException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterIdToDeployIn(),
vm.getPodIdToDeployIn(),
"Unable to restart " + vm.getHostName() + " which was running on host " + hostDesc,
"The Storage is unavailable for trying to restart VM, name: "
+ vm.getHostName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
} catch (OperationTimedoutException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterIdToDeployIn(),
vm.getPodIdToDeployIn(),
"Unable to restart " + vm.getHostName() + " which was running on host " + hostDesc,
"The Storage is unavailable for trying to restart VM, name: "
+ vm.getHostName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
}
vm = _itMgr.findByIdAndType(vm.getType(), vm.getId());
work.setUpdateTime(vm.getUpdated());
work.setPreviousState(vm.getState());
return (System.currentTimeMillis() >> 10) + _restartRetryInterval;
}
@Override
public Host addTrafficMonitor(AddTrafficMonitorCmd cmd) {
long zoneId = cmd.getZoneId();
DataCenterVO zone = _dcDao.findById(zoneId);
String zoneName;
if (zone == null) {
throw new InvalidParameterValueException("Could not find zone with ID: " + zoneId);
} else {
zoneName = zone.getName();
}
List<HostVO> trafficMonitorsInZone =
_resourceMgr.listAllHostsInOneZoneByType(Host.Type.TrafficMonitor, zoneId);
if (trafficMonitorsInZone.size() != 0) {
throw new InvalidParameterValueException(
"Already added an traffic monitor in zone: " + zoneName);
}
URI uri;
try {
uri = new URI(cmd.getUrl());
} catch (Exception e) {
s_logger.debug(e);
throw new InvalidParameterValueException(e.getMessage());
}
String ipAddress = uri.getHost();
// String numRetries = params.get("numretries");
// String timeout = params.get("timeout");
TrafficSentinelResource resource = new TrafficSentinelResource();
String guid =
getTrafficMonitorGuid(zoneId, NetworkUsageResourceName.TrafficSentinel, ipAddress);
Map<String, Object> hostParams = new HashMap<String, Object>();
hostParams.put("zone", String.valueOf(zoneId));
hostParams.put("ipaddress", ipAddress);
hostParams.put("url", cmd.getUrl());
hostParams.put("inclZones", (cmd.getInclZones() != null) ? cmd.getInclZones() : _TSinclZones);
hostParams.put("exclZones", (cmd.getExclZones() != null) ? cmd.getExclZones() : _TSexclZones);
hostParams.put("guid", guid);
hostParams.put("name", guid);
try {
resource.configure(guid, hostParams);
} catch (ConfigurationException e) {
throw new CloudRuntimeException(e.getMessage());
}
Map<String, String> hostDetails = new HashMap<String, String>();
hostDetails.put("url", cmd.getUrl());
hostDetails.put("last_collection", "" + System.currentTimeMillis());
if (cmd.getInclZones() != null) {
hostDetails.put("inclZones", cmd.getInclZones());
}
if (cmd.getExclZones() != null) {
hostDetails.put("exclZones", cmd.getExclZones());
}
Host trafficMonitor =
_resourceMgr.addHost(zoneId, resource, Host.Type.TrafficMonitor, hostDetails);
return trafficMonitor;
}
public Long migrate(final HaWorkVO work) {
final long vmId = work.getInstanceId();
final VirtualMachineGuru<VMInstanceVO> mgr = findManager(work.getType());
VMInstanceVO vm = mgr.get(vmId);
if (vm == null || vm.getRemoved() != null) {
s_logger.debug("Unable to find the vm " + vmId);
return null;
}
s_logger.info("Migrating vm: " + vm.toString());
if (vm.getHostId() == null || vm.getHostId() != work.getHostId()) {
s_logger.info("VM is not longer running on the current hostId");
return null;
}
short alertType = AlertManager.ALERT_TYPE_USERVM_MIGRATE;
if (VirtualMachine.Type.DomainRouter.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_DOMAIN_ROUTER_MIGRATE;
} else if (VirtualMachine.Type.ConsoleProxy.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_CONSOLE_PROXY_MIGRATE;
}
HostVO fromHost = _hostDao.findById(vm.getHostId());
String fromHostName = ((fromHost == null) ? "unknown" : fromHost.getName());
HostVO toHost = null;
if (work.getStep() == Step.Scheduled) {
if (vm.getState() != State.Running) {
s_logger.info(
"VM's state is not ready for migration. "
+ vm.toString()
+ " State is "
+ vm.getState().toString());
return (System.currentTimeMillis() >> 10) + _migrateRetryInterval;
}
DataCenterVO dcVO = _dcDao.findById(fromHost.getDataCenterId());
HostPodVO podVO = _podDao.findById(fromHost.getPodId());
try {
toHost = mgr.prepareForMigration(vm);
if (toHost == null) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("Unable to find a host for migrating vm " + vmId);
}
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to migrate vm "
+ vm.getName()
+ " from host "
+ fromHostName
+ " in zone "
+ dcVO.getName()
+ " and pod "
+ podVO.getName(),
"Unable to find a suitable host");
}
} catch (final InsufficientCapacityException e) {
s_logger.warn("Unable to mgirate due to insufficient capacity " + vm.toString());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to migrate vm "
+ vm.getName()
+ " from host "
+ fromHostName
+ " in zone "
+ dcVO.getName()
+ " and pod "
+ podVO.getName(),
"Insufficient capacity");
} catch (final StorageUnavailableException e) {
s_logger.warn("Storage is unavailable: " + vm.toString());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to migrate vm "
+ vm.getName()
+ " from host "
+ fromHostName
+ " in zone "
+ dcVO.getName()
+ " and pod "
+ podVO.getName(),
"Storage is gone.");
}
if (toHost == null) {
_agentMgr.maintenanceFailed(vm.getHostId());
return null;
}
if (s_logger.isDebugEnabled()) {
s_logger.debug("Migrating from " + work.getHostId() + " to " + toHost.getId());
}
work.setStep(Step.Migrating);
work.setHostId(toHost.getId());
_haDao.update(work.getId(), work);
}
if (work.getStep() == Step.Migrating) {
vm = mgr.get(vmId); // let's see if anything has changed.
boolean migrated = false;
if (vm == null
|| vm.getRemoved() != null
|| vm.getHostId() == null
|| !_itMgr.stateTransitTo(vm, Event.MigrationRequested, vm.getHostId())) {
s_logger.info("Migration cancelled because state has changed: " + vm.toString());
} else {
try {
boolean isWindows =
_guestOSCategoryDao
.findById(_guestOSDao.findById(vm.getGuestOSId()).getCategoryId())
.getName()
.equalsIgnoreCase("Windows");
MigrateCommand cmd =
new MigrateCommand(vm.getInstanceName(), toHost.getPrivateIpAddress(), isWindows);
Answer answer = _agentMgr.send(fromHost.getId(), cmd);
if (answer != null && answer.getResult()) {
migrated = true;
_storageMgr.unshare(vm, fromHost);
work.setStep(Step.Investigating);
_haDao.update(work.getId(), work);
}
} catch (final AgentUnavailableException e) {
s_logger.debug("host became unavailable");
} catch (final OperationTimedoutException e) {
s_logger.debug("operation timed out");
if (e.isActive()) {
scheduleRestart(vm, true);
}
}
}
if (!migrated) {
s_logger.info("Migration was unsuccessful. Cleaning up: " + vm.toString());
DataCenterVO dcVO = _dcDao.findById(vm.getDataCenterId());
HostPodVO podVO = _podDao.findById(vm.getPodId());
_alertMgr.sendAlert(
alertType,
fromHost.getDataCenterId(),
fromHost.getPodId(),
"Unable to migrate vm "
+ vm.getName()
+ " from host "
+ fromHost.getName()
+ " in zone "
+ dcVO.getName()
+ " and pod "
+ podVO.getName(),
"Migrate Command failed. Please check logs.");
_itMgr.stateTransitTo(vm, Event.MigrationFailedOnSource, toHost.getId());
_agentMgr.maintenanceFailed(vm.getHostId());
Command cleanup = mgr.cleanup(vm, null);
_agentMgr.easySend(toHost.getId(), cleanup);
_storageMgr.unshare(vm, toHost);
return null;
}
}
if (toHost == null) {
toHost = _hostDao.findById(work.getHostId());
}
DataCenterVO dcVO = _dcDao.findById(toHost.getDataCenterId());
HostPodVO podVO = _podDao.findById(toHost.getPodId());
try {
if (!mgr.completeMigration(vm, toHost)) {
_alertMgr.sendAlert(
alertType,
toHost.getDataCenterId(),
toHost.getPodId(),
"Unable to migrate "
+ vmId
+ " to host "
+ toHost.getName()
+ " in zone "
+ dcVO.getName()
+ " and pod "
+ podVO.getName(),
"Migration not completed");
s_logger.warn("Unable to complete migration: " + vm.toString());
} else {
s_logger.info("Migration is complete: " + vm.toString());
}
return null;
} catch (final AgentUnavailableException e) {
s_logger.warn("Agent is unavailable for " + vm.toString());
} catch (final OperationTimedoutException e) {
s_logger.warn("Operation timed outfor " + vm.toString());
}
_itMgr.stateTransitTo(vm, Event.MigrationFailedOnDest, toHost.getId());
return (System.currentTimeMillis() >> 10) + _migrateRetryInterval;
}
/**
* compareState does as its name suggests and compares the states between management server and
* agent. It returns whether something should be cleaned up
*/
protected Command compareState(VMInstanceVO vm, final AgentVmInfo info, final boolean fullSync) {
State agentState = info.state;
final String agentName = info.name;
final State serverState = vm.getState();
final String serverName = vm.getName();
Command command = null;
if (s_logger.isDebugEnabled()) {
s_logger.debug(
"VM "
+ serverName
+ ": server state = "
+ serverState.toString()
+ " and agent state = "
+ agentState.toString());
}
if (agentState == State.Error) {
agentState = State.Stopped;
short alertType = AlertManager.ALERT_TYPE_USERVM;
if (VirtualMachine.Type.DomainRouter.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_DOMAIN_ROUTER;
} else if (VirtualMachine.Type.ConsoleProxy.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_CONSOLE_PROXY;
}
HostPodVO podVO = _podDao.findById(vm.getPodId());
DataCenterVO dcVO = _dcDao.findById(vm.getDataCenterId());
HostVO hostVO = _hostDao.findById(vm.getHostId());
String hostDesc =
"name: "
+ hostVO.getName()
+ " (id:"
+ hostVO.getId()
+ "), availability zone: "
+ dcVO.getName()
+ ", pod: "
+ podVO.getName();
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"VM (name: "
+ vm.getName()
+ ", id: "
+ vm.getId()
+ ") stopped on host "
+ hostDesc
+ " due to storage failure",
"Virtual Machine "
+ vm.getName()
+ " (id: "
+ vm.getId()
+ ") running on host ["
+ vm.getHostId()
+ "] stopped due to storage failure.");
}
if (serverState == State.Migrating) {
s_logger.debug("Skipping vm in migrating state: " + vm.toString());
return null;
}
if (agentState == serverState) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("Both states are " + agentState.toString() + " for " + serverName);
}
assert (agentState == State.Stopped || agentState == State.Running)
: "If the states we send up is changed, this must be changed.";
_itMgr.stateTransitTo(
vm,
agentState == State.Stopped
? VirtualMachine.Event.AgentReportStopped
: VirtualMachine.Event.AgentReportRunning,
vm.getHostId());
if (agentState == State.Stopped) {
s_logger.debug("State matches but the agent said stopped so let's send a cleanup anyways.");
return info.mgr.cleanup(vm, agentName);
}
return null;
}
if (agentState == State.Stopped) {
// This state means the VM on the agent was detected previously
// and now is gone. This is slightly different than if the VM
// was never completed but we still send down a Stop Command
// to ensure there's cleanup.
if (serverState == State.Running) {
// Our records showed that it should be running so let's restart it.
vm = info.mgr.get(vm.getId());
scheduleRestart(vm, false);
command = info.mgr.cleanup(vm, agentName);
} else if (serverState == State.Stopping) {
if (fullSync) {
s_logger.debug("VM is in stopping state on full sync. Updating the status to stopped");
vm = info.mgr.get(vm.getId());
info.mgr.completeStopCommand(vm);
command = info.mgr.cleanup(vm, agentName);
} else {
s_logger.debug("Ignoring VM in stopping mode: " + vm.getName());
}
} else if (serverState == State.Starting) {
s_logger.debug("Ignoring VM in starting mode: " + vm.getName());
} else {
s_logger.debug("Sending cleanup to a stopped vm: " + agentName);
_itMgr.stateTransitTo(vm, VirtualMachine.Event.AgentReportStopped, null);
command = info.mgr.cleanup(vm, agentName);
}
} else if (agentState == State.Running) {
if (serverState == State.Starting) {
if (fullSync) {
s_logger.debug("VM state is starting on full sync so updating it to running");
vm = info.mgr.get(vm.getId());
info.mgr.completeStartCommand(vm);
}
} else if (serverState == State.Stopping) {
if (fullSync) {
s_logger.debug("VM state is in stopping on fullsync so resend stop.");
vm = info.mgr.get(vm.getId());
info.mgr.completeStopCommand(vm);
command = info.mgr.cleanup(vm, agentName);
} else {
s_logger.debug("VM is in stopping state so no action.");
}
} else if (serverState == State.Destroyed
|| serverState == State.Stopped
|| serverState == State.Expunging) {
s_logger.debug("VM state is in stopped so stopping it on the agent");
vm = info.mgr.get(vm.getId());
command = info.mgr.cleanup(vm, agentName);
} else {
_itMgr.stateTransitTo(vm, VirtualMachine.Event.AgentReportRunning, vm.getHostId());
}
} /*else if (agentState == State.Unknown) {
if (serverState == State.Running) {
if (fullSync) {
vm = info.handler.get(vm.getId());
}
scheduleRestart(vm, false);
} else if (serverState == State.Starting) {
if (fullSync) {
vm = info.handler.get(vm.getId());
}
scheduleRestart(vm, false);
} else if (serverState == State.Stopping) {
if (fullSync) {
s_logger.debug("VM state is stopping in full sync. Resending stop");
command = info.handler.cleanup(vm, agentName);
}
}
}*/
return command;
}
protected Long restart(final HaWorkVO work) {
final long vmId = work.getInstanceId();
final VirtualMachineGuru<VMInstanceVO> mgr = findManager(work.getType());
if (mgr == null) {
s_logger.warn(
"Unable to find a handler for " + work.getType().toString() + ", throwing out " + vmId);
return null;
}
VMInstanceVO vm = mgr.get(vmId);
if (vm == null) {
s_logger.info("Unable to find vm: " + vmId);
return null;
}
s_logger.info("HA on " + vm.toString());
if (vm.getState() != work.getPreviousState() || vm.getUpdated() != work.getUpdateTime()) {
s_logger.info(
"VM "
+ vm.toString()
+ " has been changed. Current State = "
+ vm.getState()
+ " Previous State = "
+ work.getPreviousState()
+ " last updated = "
+ vm.getUpdated()
+ " previous updated = "
+ work.getUpdateTime());
return null;
}
final HostVO host = _hostDao.findById(work.getHostId());
DataCenterVO dcVO = _dcDao.findById(host.getDataCenterId());
HostPodVO podVO = _podDao.findById(host.getPodId());
String hostDesc =
"name: "
+ host.getName()
+ "(id:"
+ host.getId()
+ "), availability zone: "
+ dcVO.getName()
+ ", pod: "
+ podVO.getName();
short alertType = AlertManager.ALERT_TYPE_USERVM;
if (VirtualMachine.Type.DomainRouter.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_DOMAIN_ROUTER;
} else if (VirtualMachine.Type.ConsoleProxy.equals(vm.getType())) {
alertType = AlertManager.ALERT_TYPE_CONSOLE_PROXY;
}
Boolean alive = null;
if (work.getStep() == Step.Investigating) {
if (vm.getHostId() == null || vm.getHostId() != work.getHostId()) {
s_logger.info("VM " + vm.toString() + " is now no longer on host " + work.getHostId());
if (vm.getState() == State.Starting && vm.getUpdated() == work.getUpdateTime()) {
_itMgr.stateTransitTo(vm, Event.AgentReportStopped, null);
}
return null;
}
Enumeration<Investigator> en = _investigators.enumeration();
Investigator investigator = null;
while (en.hasMoreElements()) {
investigator = en.nextElement();
alive = investigator.isVmAlive(vm, host);
if (alive != null) {
s_logger.debug(
investigator.getName() + " found VM " + vm.getName() + "to be alive? " + alive);
break;
}
}
if (alive != null && alive) {
s_logger.debug("VM " + vm.getName() + " is found to be alive by " + investigator.getName());
if (host.getStatus() == Status.Up) {
compareState(vm, new AgentVmInfo(vm.getInstanceName(), mgr, State.Running), false);
return null;
} else {
s_logger.debug("Rescheduling because the host is not up but the vm is alive");
return (System.currentTimeMillis() >> 10) + _investigateRetryInterval;
}
}
boolean fenced = false;
if (alive == null || !alive) {
fenced = true;
s_logger.debug("Fencing off VM that we don't know the state of");
Enumeration<FenceBuilder> enfb = _fenceBuilders.enumeration();
while (enfb.hasMoreElements()) {
final FenceBuilder fb = enfb.nextElement();
Boolean result = fb.fenceOff(vm, host);
if (result != null && !result) {
fenced = false;
}
}
}
if (alive == null && !fenced) {
s_logger.debug("We were unable to fence off the VM " + vm.toString());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to restart " + vm.getName() + " which was running on host " + hostDesc,
"Insufficient capacity to restart VM, name: "
+ vm.getName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
return (System.currentTimeMillis() >> 10) + _restartRetryInterval;
}
mgr.completeStopCommand(vm);
work.setStep(Step.Scheduled);
_haDao.update(work.getId(), work);
}
// send an alert for VMs that stop unexpectedly
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"VM (name: "
+ vm.getName()
+ ", id: "
+ vmId
+ ") stopped unexpectedly on host "
+ hostDesc,
"Virtual Machine "
+ vm.getName()
+ " (id: "
+ vm.getId()
+ ") running on host ["
+ hostDesc
+ "] stopped unexpectedly.");
vm = mgr.get(vm.getId());
if (!_forceHA && !vm.isHaEnabled()) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("VM is not HA enabled so we're done.");
}
return null; // VM doesn't require HA
}
if (!_storageMgr.canVmRestartOnAnotherServer(vm.getId())) {
if (s_logger.isDebugEnabled()) {
s_logger.debug("VM can not restart on another server.");
}
return null;
}
if (work.getTimesTried() > _maxRetries) {
s_logger.warn("Retried to max times so deleting: " + vmId);
return null;
}
try {
VMInstanceVO started = mgr.start(vm.getId(), 0);
if (started != null) {
s_logger.info("VM is now restarted: " + vmId + " on " + started.getHostId());
return null;
}
if (s_logger.isDebugEnabled()) {
s_logger.debug(
"Rescheduling VM " + vm.toString() + " to try again in " + _restartRetryInterval);
}
vm = mgr.get(vm.getId());
work.setUpdateTime(vm.getUpdated());
work.setPreviousState(vm.getState());
return (System.currentTimeMillis() >> 10) + _restartRetryInterval;
} catch (final InsufficientCapacityException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to restart " + vm.getName() + " which was running on host " + hostDesc,
"Insufficient capacity to restart VM, name: "
+ vm.getName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
return null;
} catch (final StorageUnavailableException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to restart " + vm.getName() + " which was running on host " + hostDesc,
"The Storage is unavailable for trying to restart VM, name: "
+ vm.getName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
return null;
} catch (ConcurrentOperationException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to restart " + vm.getName() + " which was running on host " + hostDesc,
"The Storage is unavailable for trying to restart VM, name: "
+ vm.getName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
return null;
} catch (ExecutionException e) {
s_logger.warn("Unable to restart " + vm.toString() + " due to " + e.getMessage());
_alertMgr.sendAlert(
alertType,
vm.getDataCenterId(),
vm.getPodId(),
"Unable to restart " + vm.getName() + " which was running on host " + hostDesc,
"The Storage is unavailable for trying to restart VM, name: "
+ vm.getName()
+ ", id: "
+ vmId
+ " which was running on host "
+ hostDesc);
return null;
}
}
private void generateEmailAlert(
DataCenterVO dc,
HostPodVO pod,
ClusterVO cluster,
double totalCapacity,
double usedCapacity,
short capacityType) {
String msgSubject = null;
String msgContent = null;
String totalStr;
String usedStr;
String pctStr = formatPercent(usedCapacity / totalCapacity);
short alertType = -1;
Long podId = pod == null ? null : pod.getId();
Long clusterId = cluster == null ? null : cluster.getId();
switch (capacityType) {
// Cluster Level
case CapacityVO.CAPACITY_TYPE_MEMORY:
msgSubject =
"System Alert: Low Available Memory in cluster "
+ cluster.getName()
+ " pod "
+ pod.getName()
+ " of availablity zone "
+ dc.getName();
totalStr = formatBytesToMegabytes(totalCapacity);
usedStr = formatBytesToMegabytes(usedCapacity);
msgContent =
"System memory is low, total: "
+ totalStr
+ " MB, used: "
+ usedStr
+ " MB ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_MEMORY;
break;
case CapacityVO.CAPACITY_TYPE_CPU:
msgSubject =
"System Alert: Low Unallocated CPU in cluster "
+ cluster.getName()
+ " pod "
+ pod.getName()
+ " of availablity zone "
+ dc.getName();
totalStr = _dfWhole.format(totalCapacity);
usedStr = _dfWhole.format(usedCapacity);
msgContent =
"Unallocated CPU is low, total: "
+ totalStr
+ " Mhz, used: "
+ usedStr
+ " Mhz ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_CPU;
break;
case CapacityVO.CAPACITY_TYPE_STORAGE:
msgSubject =
"System Alert: Low Available Storage in cluster "
+ cluster.getName()
+ " pod "
+ pod.getName()
+ " of availablity zone "
+ dc.getName();
totalStr = formatBytesToMegabytes(totalCapacity);
usedStr = formatBytesToMegabytes(usedCapacity);
msgContent =
"Available storage space is low, total: "
+ totalStr
+ " MB, used: "
+ usedStr
+ " MB ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_STORAGE;
break;
case CapacityVO.CAPACITY_TYPE_STORAGE_ALLOCATED:
msgSubject =
"System Alert: Remaining unallocated Storage is low in cluster "
+ cluster.getName()
+ " pod "
+ pod.getName()
+ " of availablity zone "
+ dc.getName();
totalStr = formatBytesToMegabytes(totalCapacity);
usedStr = formatBytesToMegabytes(usedCapacity);
msgContent =
"Unallocated storage space is low, total: "
+ totalStr
+ " MB, allocated: "
+ usedStr
+ " MB ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_STORAGE_ALLOCATED;
break;
case CapacityVO.CAPACITY_TYPE_LOCAL_STORAGE:
msgSubject =
"System Alert: Remaining unallocated Local Storage is low in cluster "
+ cluster.getName()
+ " pod "
+ pod.getName()
+ " of availablity zone "
+ dc.getName();
totalStr = formatBytesToMegabytes(totalCapacity);
usedStr = formatBytesToMegabytes(usedCapacity);
msgContent =
"Unallocated storage space is low, total: "
+ totalStr
+ " MB, allocated: "
+ usedStr
+ " MB ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_LOCAL_STORAGE;
break;
// Pod Level
case CapacityVO.CAPACITY_TYPE_PRIVATE_IP:
msgSubject =
"System Alert: Number of unallocated private IPs is low in pod "
+ pod.getName()
+ " of availablity zone "
+ dc.getName();
totalStr = Double.toString(totalCapacity);
usedStr = Double.toString(usedCapacity);
msgContent =
"Number of unallocated private IPs is low, total: "
+ totalStr
+ ", allocated: "
+ usedStr
+ " ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_PRIVATE_IP;
break;
// Zone Level
case CapacityVO.CAPACITY_TYPE_SECONDARY_STORAGE:
msgSubject =
"System Alert: Low Available Secondary Storage in availablity zone " + dc.getName();
totalStr = formatBytesToMegabytes(totalCapacity);
usedStr = formatBytesToMegabytes(usedCapacity);
msgContent =
"Available secondary storage space is low, total: "
+ totalStr
+ " MB, used: "
+ usedStr
+ " MB ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_SECONDARY_STORAGE;
break;
case CapacityVO.CAPACITY_TYPE_VIRTUAL_NETWORK_PUBLIC_IP:
msgSubject =
"System Alert: Number of unallocated virtual network public IPs is low in availablity zone "
+ dc.getName();
totalStr = Double.toString(totalCapacity);
usedStr = Double.toString(usedCapacity);
msgContent =
"Number of unallocated public IPs is low, total: "
+ totalStr
+ ", allocated: "
+ usedStr
+ " ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_VIRTUAL_NETWORK_PUBLIC_IP;
break;
case CapacityVO.CAPACITY_TYPE_DIRECT_ATTACHED_PUBLIC_IP:
msgSubject =
"System Alert: Number of unallocated direct attached public IPs is low in availablity zone "
+ dc.getName();
totalStr = Double.toString(totalCapacity);
usedStr = Double.toString(usedCapacity);
msgContent =
"Number of unallocated direct attached public IPs is low, total: "
+ totalStr
+ ", allocated: "
+ usedStr
+ " ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_DIRECT_ATTACHED_PUBLIC_IP;
break;
case CapacityVO.CAPACITY_TYPE_VLAN:
msgSubject =
"System Alert: Number of unallocated VLANs is low in availablity zone " + dc.getName();
totalStr = Double.toString(totalCapacity);
usedStr = Double.toString(usedCapacity);
msgContent =
"Number of unallocated VLANs is low, total: "
+ totalStr
+ ", allocated: "
+ usedStr
+ " ("
+ pctStr
+ "%)";
alertType = ALERT_TYPE_VLAN;
break;
}
try {
if (s_logger.isDebugEnabled()) {
s_logger.debug(msgSubject);
s_logger.debug(msgContent);
}
_emailAlert.sendAlert(alertType, dc.getId(), podId, clusterId, msgSubject, msgContent);
} catch (Exception ex) {
s_logger.error("Exception in CapacityChecker", ex);
}
}