Skip to content

kubernetes: clean up temporary provisioning scripts on the management server #13592

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
nikolauseppinger wants to merge 1 commit into
base: main
Choose a base branch
from
Open

kubernetes: clean up temporary provisioning scripts on the management server #13592

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 29 additions & 2 deletions ...c/main/java/com/cloud/kubernetes/cluster/actionworkers/KubernetesClusterActionWorker.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -761,6 +761,25 @@ protected void copyScriptFile(String nodeAddress, final int sshPort, File file,
}
}

/**
* Deletes the local temporary copies of the scripts created by {@link #retrieveScriptFiles()}
* once they have been copied to the cluster node(s). Without this, every deploy/autoscale/PV-cleanup
* action leaks a *.sh file in the management server's tmp directory.
*/
protected void cleanupScriptFiles() {
deleteScriptFileQuietly(deploySecretsScriptFile);
deleteScriptFileQuietly(deployProviderScriptFile);
deleteScriptFileQuietly(deployCsiDriverScriptFile);
deleteScriptFileQuietly(deletePvScriptFile);
deleteScriptFileQuietly(autoscaleScriptFile);
}

protected void deleteScriptFileQuietly(File file) {
if (file != null && file.exists() && !file.delete()) {
logger.debug("Failed to delete temporary Kubernetes cluster script file: {}", file.getAbsolutePath());
}
}

protected boolean taintControlNodes() {
StringBuilder commands = new StringBuilder();
List<KubernetesClusterVmMapVO> vmMapVOList = getKubernetesClusterVMMaps();
Expand Down Expand Up @@ -820,7 +839,11 @@ protected boolean deployProvider() {
if (!result.first()) {
logMessage(Level.INFO, "Provider files missing. Adding them now", null);
retrieveScriptFiles();
copyScripts(publicIpAddress, sshPort);
try {
copyScripts(publicIpAddress, sshPort);
} finally {
cleanupScriptFiles();
}

if (!createCloudStackSecret(keys)) {
logTransitStateAndThrow(Level.ERROR, String.format("Failed to setup keys for Kubernetes cluster %s",
Expand Down Expand Up @@ -857,7 +880,11 @@ protected boolean deployCsiDriver() {
if (!result.first()) {
logMessage(Level.INFO, "CSI files missing. Adding them now", null);
retrieveScriptFiles();
copyScripts(publicIpAddress, sshPort);
try {
copyScripts(publicIpAddress, sshPort);
} finally {
cleanupScriptFiles();
}

if (!createCloudStackSecret(keys)) {
logTransitStateAndThrow(Level.ERROR, String.format("Failed to setup keys for Kubernetes cluster %s",
Expand Down
34 changes: 21 additions & 13 deletions ...cloud/kubernetes/cluster/actionworkers/KubernetesClusterResourceModifierActionWorker.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -951,7 +951,11 @@ protected boolean autoscaleCluster(boolean enable, Long minSize, Long maxSize) {
if (!result.first()) {
logMessage(Level.INFO, "Autoscaling files missing. Adding them now", null);
retrieveScriptFiles();
copyScripts(publicIpAddress, sshPort);
try {
copyScripts(publicIpAddress, sshPort);
} finally {
cleanupScriptFiles();
}

if (!createCloudStackSecret(keys)) {
logTransitStateAndThrow(Level.ERROR, String.format("Failed to setup keys for Kubernetes cluster %s",
Expand Down Expand Up @@ -1003,19 +1007,23 @@ public boolean deletePVsWithReclaimPolicyDelete() {
if (Boolean.FALSE.equals(result.first())) {
logMessage(Level.INFO, "PV delete script missing. Adding it now", null);
retrieveScriptFiles();
if (deletePvScriptFile != null) {
copyScriptFile(publicIpAddress, sshPort, deletePvScriptFile, deletePvScriptFilename);
logMessage(Level.INFO, "Executing PV deletion script (this may take several minutes)...", null);
result = SshHelper.sshExecute(publicIpAddress, sshPort, getControlNodeLoginUser(),
pkFile, null, command, 10000, 10000, 600000); // 10 minute timeout
if (Boolean.FALSE.equals(result.first())) {
logMessage(Level.ERROR, "PV deletion script failed: " + result.second(), null);
throw new CloudRuntimeException(result.second());
try {
if (deletePvScriptFile != null) {
copyScriptFile(publicIpAddress, sshPort, deletePvScriptFile, deletePvScriptFilename);
logMessage(Level.INFO, "Executing PV deletion script (this may take several minutes)...", null);
result = SshHelper.sshExecute(publicIpAddress, sshPort, getControlNodeLoginUser(),
pkFile, null, command, 10000, 10000, 600000); // 10 minute timeout
if (Boolean.FALSE.equals(result.first())) {
logMessage(Level.ERROR, "PV deletion script failed: " + result.second(), null);
throw new CloudRuntimeException(result.second());
}
logMessage(Level.INFO, "PV deletion script completed successfully", null);
} else {
logMessage(Level.WARN, "PV delete script file not found in resources, skipping PV deletion", null);
return false;
}
logMessage(Level.INFO, "PV deletion script completed successfully", null);
} else {
logMessage(Level.WARN, "PV delete script file not found in resources, skipping PV deletion", null);
return false;
} finally {
cleanupScriptFiles();
}
} else {
logMessage(Level.INFO, "PV deletion script completed successfully", null);
Expand Down
12 changes: 11 additions & 1 deletion .../main/java/com/cloud/kubernetes/cluster/actionworkers/KubernetesClusterUpgradeWorker.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,6 +63,12 @@ protected void retrieveScriptFiles() {
upgradeScriptFile = retrieveScriptFile(upgradeScriptFilename);
}

@Override
protected void cleanupScriptFiles() {
super.cleanupScriptFiles();
deleteScriptFileQuietly(upgradeScriptFile);
}

private Pair<Boolean, String> runInstallScriptOnVM(final UserVm vm, final int index) throws Exception {
int nodeSshPort = sshPort == 22 ? sshPort : sshPort + index;
String nodeAddress = (index > 0 && sshPort == 22) ? vm.getPrivateIpAddress() : publicIpAddress;
Expand Down Expand Up @@ -176,7 +182,11 @@ public boolean upgradeCluster() throws CloudRuntimeException {
retrieveScriptFiles();
stateTransitTo(kubernetesCluster.getId(), KubernetesCluster.Event.UpgradeRequested);
attachIsoKubernetesVMs(clusterVMs, upgradeVersion);
upgradeKubernetesClusterNodes();
try {
upgradeKubernetesClusterNodes();
} finally {
cleanupScriptFiles();
}
detachIsoKubernetesVMs(clusterVMs);
KubernetesClusterVO kubernetesClusterVO = kubernetesClusterDao.findById(kubernetesCluster.getId());
kubernetesClusterVO.setKubernetesVersionId(upgradeVersion.getId());
Expand Down
45 changes: 45 additions & 0 deletions ...st/java/com/cloud/kubernetes/cluster/actionworkers/KubernetesClusterActionWorkerTest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,7 @@
// under the License.
package com.cloud.kubernetes.cluster.actionworkers;

import java.io.File;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
Expand Down Expand Up @@ -229,4 +230,48 @@ public void testGetMergedAffinityGroupIdsExplicitDedicationAlreadyInList() {
Assert.assertTrue(result.contains(99L));
Assert.assertTrue(result.contains(2L));
}

@Test
public void testCleanupScriptFilesDeletesAllTempFiles() throws Exception {
actionWorker.deploySecretsScriptFile = File.createTempFile("deploy-cloudstack-secret", ".sh");
actionWorker.deployProviderScriptFile = File.createTempFile("deploy-provider", ".sh");
actionWorker.deployCsiDriverScriptFile = File.createTempFile("deploy-csi-driver", ".sh");
actionWorker.deletePvScriptFile = File.createTempFile("delete-pv-reclaimpolicy-delete", ".sh");
actionWorker.autoscaleScriptFile = File.createTempFile("autoscale-kube-cluster", ".sh");

List<File> files = Arrays.asList(actionWorker.deploySecretsScriptFile, actionWorker.deployProviderScriptFile,
actionWorker.deployCsiDriverScriptFile, actionWorker.deletePvScriptFile, actionWorker.autoscaleScriptFile);
for (File f : files) {
Assert.assertTrue(f.exists());
}

actionWorker.cleanupScriptFiles();

for (File f : files) {
Assert.assertFalse("Temporary script file should have been deleted: " + f.getAbsolutePath(), f.exists());
}
}

@Test
public void testDeleteScriptFileQuietlyHandlesNullAndMissingFiles() throws Exception {
// null must not throw
actionWorker.deleteScriptFileQuietly(null);

// a File that does not exist must not throw
File missing = new File(System.getProperty("java.io.tmpdir"), "cks-nonexistent-" + UUID.randomUUID() + ".sh");
Assert.assertFalse(missing.exists());
actionWorker.deleteScriptFileQuietly(missing);

// an existing file is deleted
File present = File.createTempFile("cks-present", ".sh");
Assert.assertTrue(present.exists());
actionWorker.deleteScriptFileQuietly(present);
Assert.assertFalse(present.exists());
}

@Test
public void testCleanupScriptFilesWithNullFieldsDoesNotThrow() {
// No retrieveScriptFiles() was called, so all file fields are null.
actionWorker.cleanupScriptFiles();
}
}