Skip to content
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.

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

Stop Kafka Service Before Running Broker Recovery Process #309

Merged
merged 5 commits into from
Nov 26, 2024
Merged
Changes from 4 commits
Commits
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
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
import com.pinterest.orion.core.actions.alert.AlertLevel;
import com.pinterest.orion.core.actions.alert.AlertMessage;
import com.pinterest.orion.core.actions.aws.ReplaceEC2InstanceAction;
import com.pinterest.orion.core.actions.generic.GenericActions.ServiceStopAction;
import com.pinterest.orion.core.actions.generic.NodeAction;
import com.pinterest.orion.core.actions.generic.ServiceStabilityCheckAction;
import com.pinterest.orion.server.OrionServer;
Expand All @@ -31,6 +32,7 @@
import java.net.InetSocketAddress;
import java.net.Socket;
import java.util.Map;
import java.util.concurrent.TimeUnit;
import java.util.logging.Logger;

public class BrokerRecoveryAction extends NodeAction {
Expand All @@ -42,6 +44,7 @@ public class BrokerRecoveryAction extends NodeAction {
private static final int retryIntervalMilliseconds = 1000;

public static final String ATTR_TRY_TO_RESTART_KEY = "try_restart";
public static final String ATTR_STOP_SERVICE_BEFORE_REPLACEMENT = "stop_service_before_replacement";
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why do we need this? We should always do replacements by stopping the broker first, so no need to make it configurable in my opinion.

public static final String ATTR_NODE_EXISTS_KEY = "node_exists";
public static final String ATTR_NONEXISTENT_HOST_KEY = "nonexistent_host";
public static final String CONF_DRY_RUN_REPLACEMENT_KEY = "dry_run";
Expand Down Expand Up @@ -126,6 +129,37 @@ public void runAction() {

getResult().appendOut("Start replacing broker");

// Try to gracefully shutdown service before replacing broker.
// This can make sure the broker is not the leader of any partition.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: replace this comment with

// This ensures partition leaderships are moved away from the broker before the replacement process starts.

boolean stopServiceBeforeReplacement = true;
if (containsAttribute(ATTR_STOP_SERVICE_BEFORE_REPLACEMENT)) {
stopServiceBeforeReplacement = getAttribute(ATTR_STOP_SERVICE_BEFORE_REPLACEMENT).getValue();
}
if (stopServiceBeforeReplacement) {
OrionServer.METRICS.counter(metricPrefix.resolve(ATTR_STOP_SERVICE_BEFORE_REPLACEMENT)).inc();
try {
ServiceStopAction stopServiceAction = new ServiceStopAction();
stopServiceAction.copyAttributeFrom(this, OrionConstants.NODE_ID);
getEngine().dispatchChild(this, stopServiceAction);
getResult().appendOut("Attempt to stop service before broker replacement.");
try {
stopServiceAction.get(30, TimeUnit.SECONDS);
} catch (Exception e) {
getResult().appendErr(
"Unable to stop service before broker replacement. Will keep moving forward. Error message:"
+ e.getMessage());
}
} catch (Exception e) {
getResult().appendErr(
"Encounter unknown error when stopping service. Will keep moving forward. Error message:"
+ e.getMessage());
}
} else {
getResult().appendOut(
"Skip stopping service before broker replacement. " +
"The stop_service_before_replacement configuration is set to false.");
}

try {
// if dry run skip dispatching the replacement action and report success, but still set the cooldown flag for the cluster
if (!isDryRun) {
Expand Down