You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@pulsar.apache.org by "heesung-sn (via GitHub)" <gi...@apache.org> on 2023/05/13 00:24:21 UTC
[GitHub] [pulsar] heesung-sn opened a new pull request, #20315: [improve][broker] Gracefully shut down load balancer extension
heesung-sn opened a new pull request, #20315:
URL: https://github.com/apache/pulsar/pull/20315
<!--
### Contribution Checklist
- PR title format should be *[type][component] summary*. For details, see *[Guideline - Pulsar PR Naming Convention](https://pulsar.apache.org/contribute/develop-semantic-title/)*.
- Fill out the template below to describe the changes contributed by the pull request. That will give reviewers the context they need to do the review.
- Each pull request should address only one issue, not mix up code from multiple issues.
- Each commit in the pull request has a meaningful commit message
- Once all items of the checklist are addressed, remove the above text and this checklist, leaving only the filled out template below.
-->
<!-- If the PR belongs to a PIP, please add the PIP link here -->
PIP: https://github.com/apache/pulsar/issues/16691
<!-- Details of when a PIP is required and how the PIP process work, please see: https://github.com/apache/pulsar/blob/master/wiki/proposals/PIP.md -->
### Motivation
Load balancer extension needs to shut down gracefully, especially when shutting down the leader broker. If the leader broker closes the leader election service too late, the service unit lookups to the leader broker will fail during the shutdown. This could delay client reconnection time.
Lookup failure logs
```
(shutdown starts)
pulsar-broker-8 pulsar-broker 2023-04-22T00:19:52,630+0000 [pulsar-service-shutdown] INFO org.apache.pulsar.broker.PulsarService - Closing PulsarService
pulsar-broker-3 pulsar-broker 2023-04-22T00:19:52,690+0000 [pulsar-io-4-5] INFO org.apache.pulsar.client.impl.ConnectionHandler - [persistent://pulsar/system/loadbalancer-service-unit-state] [pulsar-18-9] Closed connection [id: 0x907e74b0, L:/10.0.13.6:35838 ! R:pulsar-broker-8.pulsar-broker.pulsar.svc.cluster.local/10.0.18.18:6650] -- Will try again in 0.1 s
pulsar-broker-3 pulsar-broker 2023-04-22T00:19:52,691+0000 [pulsar-io-4-5] INFO org.apache.pulsar.client.impl.ConnectionHandler - [persistent://pulsar/system/loadbalancer-service-unit-state] [reader-30e6b40dd9] Closed connection [id: 0x907e74b0, L:/10.0.13.6:35838 ! R:pulsar-broker-8.pulsar-broker.pulsar.svc.cluster.local/10.0.18.18:6650] -- Will try again in 0.1 s
(znode is gone)
pulsar-broker-8 pulsar-broker 2023-04-22T00:19:52,652+0000 [pulsar-load-manager-1-1] INFO org.apache.pulsar.broker.loadbalance.extensions.channel.ServiceUnitStateChannelImpl - BrokerRegistry detected the broker:pulsar-broker-8.pulsar-broker.pulsar.svc.cluster.local:8080 registry has been deleted.
(lookup failure)
org.apache.pulsar.client.impl.ConnectionHandler - [persistent://pulsar/system/loadbalancer-service-unit-state] [reader-30e6b40dd9] Could not get connection to broker: org.apache.pulsar.client.api.PulsarClientException$ConnectException: Disconnected from server at pulsar-broker-3.pulsar-broker.pulsar.svc.cluster.local/10.0.13.6:6650 -- Will try again in 0.193 s
pulsar-broker-3 pulsar-broker 2023-04-22T00:19:52,827+0000 [main-EventThread] INFO org.apache.pulsar.broker.lookup.TopicLookupBase - Failed to lookup null for topic persistent://pulsar/system/loadbalancer-service-unit-state with error Failed to look up a broker registry:pulsar-broker-8.pulsar-broker.pulsar.svc.cluster.local:8080 for bundle:pulsar/system/0x40000000_0x50000000
pulsar-broker-3 pulsar-broker 2023-04-22T00:19:52,827+0000 [main-EventThread] INFO org.apache.pulsar.broker.lookup.TopicLookupBase - Failed to lookup null for topic persistent://pulsar/system/loadbalancer-service-unit-state with error Failed to look up a broker registry:pulsar-broker-8.pulsar-broker.pulsar.svc.cluster.local:8080 for bundle:pulsar/system/0x40000000_0x50000000
(leader election service has been closed)
pulsar-broker-3 pulsar-broker 2023-04-22T00:19:55,569+0000 [pulsar-load-manager-1-1] INFO org.apache.pulsar.broker.loadbalance.extensions.ExtensibleLoadManagerImpl - This broker:pulsar-broker-3.pulsar-broker.pulsar.svc.cluster.local:8080 plays the leader now.
```
<!-- Explain here the context, and why you're making that change. What is the problem you're trying to solve. -->
### Modifications
During the shutdown flow, when calling loadbalancer.disableBroker(), the load balancer extension gracefully gives up the owned bundles to other brokers. After that, it closes the leader election service and removes its register in the metadata store.
<!-- Describe the modifications you've done. -->
### Verifying this change
- [x] Make sure that the change passes the CI checks.
This change is already covered by existing tests, and I added a new test case to cover this diableBroker logic.
### Does this pull request potentially affect one of the following parts:
<!-- DO NOT REMOVE THIS SECTION. CHECK THE PROPER BOX ONLY. -->
*If the box was checked, please highlight the changes*
- [ ] Dependencies (add or upgrade a dependency)
- [ ] The public API
- [ ] The schema
- [ ] The default values of configurations
- [ ] The threading model
- [ ] The binary protocol
- [ ] The REST endpoints
- [ ] The admin CLI options
- [ ] The metrics
- [ ] Anything that affects deployment
### Documentation
<!-- DO NOT REMOVE THIS SECTION. CHECK THE PROPER BOX ONLY. -->
- [ ] `doc` <!-- Your PR contains doc changes. -->
- [ ] `doc-required` <!-- Your PR changes impact docs and you will update later -->
- [x] `doc-not-needed` <!-- Your PR changes do not impact docs -->
- [ ] `doc-complete` <!-- Docs have been already added -->
### Matching PR in forked repository
PR in forked repository: <!-- ENTER URL HERE -->
<!--
After opening this PR, the build in apache/pulsar will fail and instructions will
be provided for opening a PR in the PR author's forked repository.
apache/pulsar pull requests should be first tested in your own fork since the
apache/pulsar CI based on GitHub Actions has constrained resources and quota.
GitHub Actions provides separate quota for pull requests that are executed in
a forked repository.
The tests will be run in the forked repository until all PR review comments have
been handled, the tests pass and the PR is approved by a reviewer.
-->
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] Demogorgon314 commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "Demogorgon314 (via GitHub)" <gi...@apache.org>.
Demogorgon314 commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1192960768
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/ExtensibleLoadManagerImpl.java:
##########
@@ -380,12 +380,22 @@ public CompletableFuture<Optional<BrokerLookupData>> assign(Optional<ServiceUnit
}
public CompletableFuture<Optional<String>> selectAsync(ServiceUnitId bundle) {
+ return selectAsync(bundle, Optional.empty());
+ }
+
+ public CompletableFuture<Optional<String>> selectAsync(ServiceUnitId bundle,
+ Optional<Set<String>> excludeBrokerSet) {
Review Comment:
+1
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lifepuzzlefun commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "lifepuzzlefun (via GitHub)" <gi...@apache.org>.
lifepuzzlefun commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1192907177
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/ExtensibleLoadManagerImpl.java:
##########
@@ -380,12 +380,22 @@ public CompletableFuture<Optional<BrokerLookupData>> assign(Optional<ServiceUnit
}
public CompletableFuture<Optional<String>> selectAsync(ServiceUnitId bundle) {
+ return selectAsync(bundle, Optional.empty());
+ }
+
+ public CompletableFuture<Optional<String>> selectAsync(ServiceUnitId bundle,
+ Optional<Set<String>> excludeBrokerSet) {
Review Comment:
maybe we can just use empty set like `Collections.emptySet()` to replace `Optional`
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/ExtensibleLoadManagerImpl.java:
##########
@@ -685,4 +695,10 @@ private void monitor() {
log.error("Failed to get the channel ownership.", e);
}
}
+
+ public void disableBroker() throws Exception {
+ serviceUnitStateChannel.cleanOwnerships();
+ leaderElectionService.close();
Review Comment:
I think we need keep the leaderElectionService open.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] heesung-sn commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "heesung-sn (via GitHub)" <gi...@apache.org>.
heesung-sn commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1198130991
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/channel/ServiceUnitStateChannelImpl.java:
##########
@@ -1140,26 +1152,69 @@ private void overrideOwnership(String serviceUnit, ServiceUnitStateData orphanDa
}
}
+ protected void waitForCleanups(boolean excludeSystemTopics, int timeoutInMillis) {
+ long started = System.currentTimeMillis();
+ while (System.currentTimeMillis() - started < timeoutInMillis) {
+ boolean cleaned = true;
+ for (var etr : tableview.entrySet()) {
+ var serviceUnit = etr.getKey();
+ var data = etr.getValue();
+
+ if (excludeSystemTopics && serviceUnit.startsWith(SYSTEM_NAMESPACE.toString())) {
+ continue;
+ }
+
+ if (data.state() == Owned && data.dstBroker().equals(lookupServiceAddress)) {
+ cleaned = false;
+ break;
+ }
+ }
+ if (cleaned) {
+ try {
+ MILLISECONDS.sleep(OWNERSHIP_CLEAN_UP_CONVERGENCE_DELAY_IN_MILLIS);
Review Comment:
I added a comment on line 1231 to explain this.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] gaoran10 commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "gaoran10 (via GitHub)" <gi...@apache.org>.
gaoran10 commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1197391067
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/channel/ServiceUnitStateChannelImpl.java:
##########
@@ -1140,26 +1152,69 @@ private void overrideOwnership(String serviceUnit, ServiceUnitStateData orphanDa
}
}
+ protected void waitForCleanups(boolean excludeSystemTopics, int timeoutInMillis) {
+ long started = System.currentTimeMillis();
+ while (System.currentTimeMillis() - started < timeoutInMillis) {
+ boolean cleaned = true;
+ for (var etr : tableview.entrySet()) {
+ var serviceUnit = etr.getKey();
+ var data = etr.getValue();
+
+ if (excludeSystemTopics && serviceUnit.startsWith(SYSTEM_NAMESPACE.toString())) {
+ continue;
+ }
+
+ if (data.state() == Owned && data.dstBroker().equals(lookupServiceAddress)) {
+ cleaned = false;
+ break;
+ }
+ }
+ if (cleaned) {
+ try {
+ MILLISECONDS.sleep(OWNERSHIP_CLEAN_UP_CONVERGENCE_DELAY_IN_MILLIS);
Review Comment:
Sorry, I'm not sure why need to sleep while all service units are all cleaned.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] Demogorgon314 merged pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "Demogorgon314 (via GitHub)" <gi...@apache.org>.
Demogorgon314 merged PR #20315:
URL: https://github.com/apache/pulsar/pull/20315
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] gaoran10 commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "gaoran10 (via GitHub)" <gi...@apache.org>.
gaoran10 commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1199602759
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/channel/ServiceUnitStateChannelImpl.java:
##########
@@ -1140,26 +1152,69 @@ private void overrideOwnership(String serviceUnit, ServiceUnitStateData orphanDa
}
}
+ protected void waitForCleanups(boolean excludeSystemTopics, int timeoutInMillis) {
+ long started = System.currentTimeMillis();
+ while (System.currentTimeMillis() - started < timeoutInMillis) {
+ boolean cleaned = true;
+ for (var etr : tableview.entrySet()) {
+ var serviceUnit = etr.getKey();
+ var data = etr.getValue();
+
+ if (excludeSystemTopics && serviceUnit.startsWith(SYSTEM_NAMESPACE.toString())) {
+ continue;
+ }
+
+ if (data.state() == Owned && data.dstBroker().equals(lookupServiceAddress)) {
+ cleaned = false;
+ break;
+ }
+ }
+ if (cleaned) {
+ try {
+ MILLISECONDS.sleep(OWNERSHIP_CLEAN_UP_CONVERGENCE_DELAY_IN_MILLIS);
Review Comment:
Ok, thanks.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] heesung-sn commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "heesung-sn (via GitHub)" <gi...@apache.org>.
heesung-sn commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1192885219
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/PulsarService.java:
##########
@@ -1316,7 +1324,11 @@ public boolean isRunning() {
* @return a reference of the current <code>LeaderElectionService</code> instance.
*/
public LeaderElectionService getLeaderElectionService() {
- return this.leaderElectionService;
+ if (ExtensibleLoadManagerImpl.isLoadManagerExtensionEnabled(config)) {
Review Comment:
@Demogorgon314 Plz confirm this change.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] heesung-sn commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "heesung-sn (via GitHub)" <gi...@apache.org>.
heesung-sn commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1198130991
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/channel/ServiceUnitStateChannelImpl.java:
##########
@@ -1140,26 +1152,69 @@ private void overrideOwnership(String serviceUnit, ServiceUnitStateData orphanDa
}
}
+ protected void waitForCleanups(boolean excludeSystemTopics, int timeoutInMillis) {
+ long started = System.currentTimeMillis();
+ while (System.currentTimeMillis() - started < timeoutInMillis) {
+ boolean cleaned = true;
+ for (var etr : tableview.entrySet()) {
+ var serviceUnit = etr.getKey();
+ var data = etr.getValue();
+
+ if (excludeSystemTopics && serviceUnit.startsWith(SYSTEM_NAMESPACE.toString())) {
+ continue;
+ }
+
+ if (data.state() == Owned && data.dstBroker().equals(lookupServiceAddress)) {
+ cleaned = false;
+ break;
+ }
+ }
+ if (cleaned) {
+ try {
+ MILLISECONDS.sleep(OWNERSHIP_CLEAN_UP_CONVERGENCE_DELAY_IN_MILLIS);
Review Comment:
This is a graceful delay for other brokers(consumers) to converge on the cleanup. Here, we can't wait for all other brokers to consume the cleanup messages. However, this delay helps the cluster to have a relatively consistent view before return.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] heesung-sn commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "heesung-sn (via GitHub)" <gi...@apache.org>.
heesung-sn commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1193002354
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/ExtensibleLoadManagerImpl.java:
##########
@@ -380,12 +380,22 @@ public CompletableFuture<Optional<BrokerLookupData>> assign(Optional<ServiceUnit
}
public CompletableFuture<Optional<String>> selectAsync(ServiceUnitId bundle) {
+ return selectAsync(bundle, Optional.empty());
+ }
+
+ public CompletableFuture<Optional<String>> selectAsync(ServiceUnitId bundle,
+ Optional<Set<String>> excludeBrokerSet) {
Review Comment:
Updated.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lifepuzzlefun commented on a diff in pull request #20315: [improve][broker] Gracefully shut down load balancer extension
Posted by "lifepuzzlefun (via GitHub)" <gi...@apache.org>.
lifepuzzlefun commented on code in PR #20315:
URL: https://github.com/apache/pulsar/pull/20315#discussion_r1192909962
##########
pulsar-broker/src/main/java/org/apache/pulsar/broker/loadbalance/extensions/ExtensibleLoadManagerImpl.java:
##########
@@ -685,4 +695,10 @@ private void monitor() {
log.error("Failed to get the channel ownership.", e);
}
}
+
+ public void disableBroker() throws Exception {
+ serviceUnitStateChannel.cleanOwnerships();
+ leaderElectionService.close();
Review Comment:
I think we need keep the leaderElectionService open.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org