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 void scheduleRestart(VMInstanceVO vm, boolean investigate) {
Long hostId = vm.getHostId();
if (hostId == null) {
try {
s_logger.debug("Found a vm that is scheduled to be restarted but has no host id: " + vm);
_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);
}
return;
}
if (vm.getHypervisorType() == HypervisorType.VMware) {
s_logger.info("Skip HA for VMware VM " + vm.getInstanceName());
return;
}
if (!investigate) {
if (s_logger.isDebugEnabled()) {
s_logger.debug(
"VM does not require investigation so I'm marking it as Stopped: " + vm.toString());
}
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;
}
if (!(_forceHA || vm.isHaEnabled())) {
String hostDesc =
"id:"
+ vm.getHostId()
+ ", availability zone id:"
+ vm.getDataCenterIdToDeployIn()
+ ", pod id:"
+ vm.getPodIdToDeployIn();
_alertMgr.sendAlert(
alertType,
vm.getDataCenterIdToDeployIn(),
vm.getPodIdToDeployIn(),
"VM (name: "
+ vm.getHostName()
+ ", id: "
+ vm.getId()
+ ") stopped unexpectedly on host "
+ hostDesc,
"Virtual Machine "
+ vm.getHostName()
+ " (id: "
+ vm.getId()
+ ") running on host ["
+ vm.getHostId()
+ "] stopped unexpectedly.");
if (s_logger.isDebugEnabled()) {
s_logger.debug("VM is not HA enabled so we're done.");
}
}
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);
}
}
List<HaWorkVO> items = _haDao.findPreviousHA(vm.getId());
int maxRetries = 0;
for (HaWorkVO item : items) {
if (maxRetries < item.getTimesTried() && !item.canScheduleNew(_timeBetweenFailures)) {
maxRetries = item.getTimesTried();
break;
}
}
HaWorkVO work =
new HaWorkVO(
vm.getId(),
vm.getType(),
WorkType.HA,
investigate ? Step.Investigating : Step.Scheduled,
hostId,
vm.getState(),
maxRetries + 1,
vm.getUpdated());
_haDao.persist(work);
if (s_logger.isInfoEnabled()) {
s_logger.info("Schedule vm for HA: " + vm);
}
wakeupWorkers();
}