Examples with AVAILABLE com.sequenceiq.cloudbreak.api.endpoint.v4.common.Status.AVAILABLE used on opensource projects

Example 1 with AVAILABLE

use of com.sequenceiq.cloudbreak.api.endpoint.v4.common.Status.AVAILABLE in project cloudbreak by hortonworks.

the class ClusterHostServiceRunner method changePrimaryGateway.

public String changePrimaryGateway(Stack stack) throws CloudbreakException {
    GatewayConfig formerPrimaryGatewayConfig = gatewayConfigService.getPrimaryGatewayConfig(stack);
    List<GatewayConfig> gatewayConfigs = gatewayConfigService.getAllGatewayConfigs(stack);
    Optional<GatewayConfig> newPrimaryCandidate = gatewayConfigs.stream().filter(gc -> !gc.isPrimary()).findFirst();
    if (newPrimaryCandidate.isPresent()) {
        GatewayConfig newPrimary = newPrimaryCandidate.get();
        Set<Node> allNodes = stackUtil.collectNodes(stack);
        try {
            hostOrchestrator.changePrimaryGateway(formerPrimaryGatewayConfig, newPrimary, gatewayConfigs, allNodes, clusterDeletionBasedModel(stack.getId(), stack.getCluster().getId()));
            return newPrimary.getHostname();
        } catch (CloudbreakOrchestratorException ex) {
            throw new CloudbreakException(ex);
        }
    } else {
        throw new CloudbreakException("Primary gateway change is not possible because there is no available node for the action");
    }
}

Example 2 with AVAILABLE

use of com.sequenceiq.cloudbreak.api.endpoint.v4.common.Status.AVAILABLE in project cloudbreak by hortonworks.

the class StopStartDownscaleDecommissionViaCMHandler method doAccept.

@Override
protected Selectable doAccept(HandlerEvent<StopStartDownscaleDecommissionViaCMRequest> event) {
    StopStartDownscaleDecommissionViaCMRequest request = event.getData();
    LOGGER.info("StopStartDownscaleDecommissionViaCMHandler for: {}, {}", event.getData().getResourceId(), event.getData());
    try {
        Stack stack = stackService.getByIdWithLists(request.getResourceId());
        Cluster cluster = stack.getCluster();
        ClusterDecomissionService clusterDecomissionService = clusterApiConnectors.getConnector(stack).clusterDecomissionService();
        Set<String> hostNames = getHostNamesForPrivateIds(request.getInstanceIdsToDecommission(), stack);
        LOGGER.debug("Attempting to decommission hosts. count={}, hostnames={}", hostNames.size(), hostNames);
        HostGroup hostGroup = hostGroupService.getByClusterIdAndName(cluster.getId(), request.getHostGroupName()).orElseThrow(NotFoundException.notFound("hostgroup", request.getHostGroupName()));
        Map<String, InstanceMetaData> hostsToRemove = clusterDecomissionService.collectHostsToRemove(hostGroup, hostNames);
        List<String> missingHostsInCm = Collections.emptyList();
        if (hostNames.size() != hostsToRemove.size()) {
            missingHostsInCm = hostNames.stream().filter(h -> !hostsToRemove.containsKey(h)).collect(Collectors.toList());
            LOGGER.info("Found fewer instances in CM to decommission, as compared to initial ask. foundCount={}, initialCount={}, missingHostsInCm={}", hostsToRemove.size(), hostNames.size(), missingHostsInCm);
        }
        // TODO CB-14929: Potentially put the nodes into maintenance mode before decommissioning?
        // TODO CB-15132: Eventually, try parsing the results of the CM decommission, and see if a partial decommission went through in the
        // timebound specified.
        Set<String> decommissionedHostNames = Collections.emptySet();
        if (hostsToRemove.size() > 0) {
            decommissionedHostNames = clusterDecomissionService.decommissionClusterNodesStopStart(hostsToRemove, POLL_FOR_10_MINUTES);
            updateInstanceStatuses(hostsToRemove, decommissionedHostNames, InstanceStatus.DECOMMISSIONED, "decommission requested for instances");
        }
        // This doesn't handle failures. It handles scenarios where CM list APIs don't have the necessary hosts available.
        List<String> allMissingHostnames = null;
        if (missingHostsInCm.size() > 0) {
            allMissingHostnames = new LinkedList<>(missingHostsInCm);
        }
        if (hostsToRemove.size() != decommissionedHostNames.size()) {
            Set<String> finalDecommissionedHostnames = decommissionedHostNames;
            List<String> additionalMissingDecommissionHostnames = hostsToRemove.keySet().stream().filter(h -> !finalDecommissionedHostnames.contains(h)).collect(Collectors.toList());
            LOGGER.info("Decommissioned fewer instances than requested. decommissionedCount={}, expectedCount={}, initialCount={}, notDecommissioned=[{}]", decommissionedHostNames.size(), hostsToRemove.size(), hostNames.size(), additionalMissingDecommissionHostnames);
            if (allMissingHostnames == null) {
                allMissingHostnames = new LinkedList<>();
            }
            allMissingHostnames.addAll(additionalMissingDecommissionHostnames);
        }
        LOGGER.info("hostsDecommissioned: count={}, hostNames={}", decommissionedHostNames.size(), decommissionedHostNames);
        if (decommissionedHostNames.size() > 0) {
            LOGGER.debug("Attempting to put decommissioned hosts into maintenance mode. count={}", decommissionedHostNames.size());
            flowMessageService.fireEventAndLog(stack.getId(), UPDATE_IN_PROGRESS.name(), CLUSTER_SCALING_STOPSTART_DOWNSCALE_ENTERINGCMMAINTMODE, String.valueOf(decommissionedHostNames.size()));
            clusterDecomissionService.enterMaintenanceMode(decommissionedHostNames);
            flowMessageService.fireEventAndLog(stack.getId(), UPDATE_IN_PROGRESS.name(), CLUSTER_SCALING_STOPSTART_DOWNSCALE_ENTEREDCMMAINTMODE, String.valueOf(decommissionedHostNames.size()));
            LOGGER.debug("Successfully put decommissioned hosts into maintenance mode. count={}", decommissionedHostNames.size());
        } else {
            LOGGER.debug("No nodes decommissioned, hence no nodes being put into maintenance mode");
        }
        return new StopStartDownscaleDecommissionViaCMResult(request, decommissionedHostNames, allMissingHostnames);
    } catch (Exception e) {
        // TODO CB-15132: This can be improved based on where and when the Exception occurred to potentially rollback certain aspects.
        // ClusterClientInitException is one which is explicitly thrown.
        String message = "Failed while attempting to decommission nodes via CM";
        LOGGER.error(message, e);
        return new StopStartDownscaleDecommissionViaCMResult(message, e, request);
    }
}

Example 3 with AVAILABLE

use of com.sequenceiq.cloudbreak.api.endpoint.v4.common.Status.AVAILABLE in project cloudbreak by hortonworks.

the class DistroXService method validate.

private void validate(DistroXV1Request request) {
    DetailedEnvironmentResponse environment = Optional.ofNullable(environmentClientService.getByName(request.getEnvironmentName())).orElseThrow(() -> new BadRequestException("No environment name provided hence unable to obtain some important data"));
    if (environment == null) {
        throw new BadRequestException(format("'%s' Environment does not exist.", request.getEnvironmentName()));
    }
    DescribeFreeIpaResponse freeipa = freeipaClientService.getByEnvironmentCrn(environment.getCrn());
    if (freeipa == null || freeipa.getAvailabilityStatus() == null || !freeipa.getAvailabilityStatus().isAvailable()) {
        throw new BadRequestException(format("If you want to provision a Data Hub then the FreeIPA instance must be running in the '%s' Environment.", environment.getName()));
    }
    Set<String> sdxCrns = platformAwareSdxConnector.listSdxCrns(environment.getName(), environment.getCrn());
    if (sdxCrns.isEmpty()) {
        throw new BadRequestException(format("Data Lake stack cannot be found for environment CRN: %s (%s)", environment.getName(), environment.getCrn()));
    }
    Set<Pair<String, StatusCheckResult>> sdxCrnsWithAvailability = platformAwareSdxConnector.listSdxCrnsWithAvailability(environment.getName(), environment.getCrn(), sdxCrns);
    if (!sdxCrnsWithAvailability.stream().map(Pair::getValue).allMatch(statusCheckResult -> StatusCheckResult.AVAILABLE.equals(statusCheckResult))) {
        throw new BadRequestException("Data Lake stacks of environment should be available.");
    }
}

Example 4 with AVAILABLE

use of com.sequenceiq.cloudbreak.api.endpoint.v4.common.Status.AVAILABLE in project cloudbreak by hortonworks.

the class SdxRuntimeUpgradeServiceTest method testCheckForUpgradeByCrnWhenDisabledAndPatchUpdatesAvailable.

@Test
@DisplayName("Test checkForUpgradeByCrn() when Runtime Upgrade is disabled and patch updates are available")
public void testCheckForUpgradeByCrnWhenDisabledAndPatchUpdatesAvailable() {
    ImageInfoV4Response currentImageInfo = new ImageInfoV4Response();
    currentImageInfo.setImageId(IMAGE_ID);
    currentImageInfo.setCreated(1);
    currentImageInfo.setComponentVersions(creatImageComponentVersions(MATCHING_TARGET_RUNTIME, MATCHING_TARGET_RUNTIME));
    ImageInfoV4Response imageInfo = new ImageInfoV4Response();
    imageInfo.setImageId(IMAGE_ID);
    imageInfo.setCreated(1);
    imageInfo.setComponentVersions(creatImageComponentVersions(MATCHING_TARGET_RUNTIME, MATCHING_TARGET_RUNTIME));
    ImageInfoV4Response imageInfo2 = new ImageInfoV4Response();
    imageInfo2.setImageId(IMAGE_ID_LAST);
    imageInfo2.setCreated(2);
    imageInfo2.setComponentVersions(creatImageComponentVersions("7.3.0", "7.3.0"));
    List<ImageInfoV4Response> candidates = List.of(imageInfo, imageInfo2);
    response.setCurrent(currentImageInfo);
    response.setUpgradeCandidates(candidates);
    SdxUpgradeResponse expectedResponse = new SdxUpgradeResponse(response.getCurrent(), candidates, response.getReason(), response.getFlowIdentifier());
    when(sdxService.getByCrn(anyString(), anyString())).thenReturn(sdxCluster);
    when(entitlementService.runtimeUpgradeEnabled(any())).thenReturn(false);
    when(stackV4Endpoint.checkForClusterUpgradeByName(anyLong(), anyString(), any(), anyString())).thenReturn(response);
    when(regionAwareInternalCrnGenerator.getInternalCrnForServiceAsString()).thenReturn("crn");
    when(regionAwareInternalCrnGeneratorFactory.iam()).thenReturn(regionAwareInternalCrnGenerator);
    ArgumentCaptor<UpgradeV4Response> upgradeV4ResponseCaptor = ArgumentCaptor.forClass(UpgradeV4Response.class);
    when(sdxUpgradeClusterConverter.upgradeResponseToSdxUpgradeResponse(upgradeV4ResponseCaptor.capture())).thenReturn(expectedResponse);
    SdxUpgradeResponse actualResponse = underTest.checkForUpgradeByCrn(USER_CRN, STACK_CRN, new SdxUpgradeRequest(), ACCOUNT_ID);
    UpgradeV4Response capturedUpgradeV4Response = upgradeV4ResponseCaptor.getValue();
    assertEquals(expectedResponse, actualResponse);
    assertTrue(StringUtils.isEmpty(capturedUpgradeV4Response.getReason()));
    assertEquals(1, capturedUpgradeV4Response.getUpgradeCandidates().size());
    assertEquals(MATCHING_TARGET_RUNTIME, capturedUpgradeV4Response.getUpgradeCandidates().get(0).getComponentVersions().getCdp());
}

Example 5 with AVAILABLE

use of com.sequenceiq.cloudbreak.api.endpoint.v4.common.Status.AVAILABLE in project cloudbreak by hortonworks.

the class DistroXRepairTests method testEphemeralDistroXMasterRepairWithTerminatedEC2Instances.

@Test(dataProvider = TEST_CONTEXT)
@UseSpotInstances
@Description(given = "there is a running Cloudbreak, and an environment with SDX and DistroX cluster in available state", when = "recovery called on the MASTER host group of DistroX cluster, where the EC2 instance had been terminated", then = "DistroX recovery should be successful, the cluster should be up and running")
public void testEphemeralDistroXMasterRepairWithTerminatedEC2Instances(TestContext testContext) {
    String distrox = resourcePropertyProvider().getName();
    List<String> actualVolumeIds = new ArrayList<>();
    List<String> expectedVolumeIds = new ArrayList<>();
    String username = testContext.getActingUserCrn().getResource();
    String sanitizedUserName = SanitizerUtil.sanitizeWorkloadUsername(username);
    testContext.given(distrox, DistroXTestDto.class).withInstanceGroupsEntity(new DistroXInstanceGroupsBuilder(testContext).defaultHostGroup().withStorageOptimizedInstancetype().build()).when(distroXTestClient.create(), key(distrox)).await(STACK_AVAILABLE).awaitForHealthyInstances().then(this::verifyMountedDisks).then((tc, testDto, client) -> {
        CloudFunctionality cloudFunctionality = tc.getCloudProvider().getCloudFunctionality();
        List<String> instancesToDelete = distroxUtil.getInstanceIds(testDto, client, MASTER.getName());
        expectedVolumeIds.addAll(cloudFunctionality.listInstanceVolumeIds(testDto.getName(), instancesToDelete));
        cloudFunctionality.deleteInstances(testDto.getName(), instancesToDelete);
        return testDto;
    }).awaitForHostGroup(MASTER.getName(), InstanceStatus.DELETED_ON_PROVIDER_SIDE).when(distroXTestClient.repair(MASTER), key(distrox)).await(STACK_AVAILABLE, key(distrox)).awaitForHealthyInstances().then(this::verifyMountedDisks).then((tc, testDto, client) -> clouderaManagerUtil.checkClouderaManagerYarnNodemanagerRoleConfigGroups(testDto, sanitizedUserName, MOCK_UMS_PASSWORD)).then((tc, testDto, client) -> {
        CloudFunctionality cloudFunctionality = tc.getCloudProvider().getCloudFunctionality();
        List<String> instanceIds = distroxUtil.getInstanceIds(testDto, client, MASTER.getName());
        actualVolumeIds.addAll(cloudFunctionality.listInstanceVolumeIds(testDto.getName(), instanceIds));
        return testDto;
    }).then((tc, testDto, client) -> VolumeUtils.compareVolumeIdsAfterRepair(testDto, actualVolumeIds, expectedVolumeIds)).validate();
}