Skip to content

[fix](group commit) group commit select backend should consider isLoadAvailable #61555

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
mymeiyi wants to merge 1 commit into
base: master
Choose a base branch
from
Open

[fix](group commit) group commit select backend should consider isLoadAvailable #61555

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
22 changes: 16 additions & 6 deletions fe/fe-core/src/main/java/org/apache/doris/load/GroupCommitManager.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -357,10 +357,7 @@ private Long getCachedBackend(String cluster, long tableId) {
return null;
}
Backend backend = Env.getCurrentSystemInfo().getBackend(backendId);
if (backend != null && backend.isAlive() && !backend.isDecommissioned()
&& (!Config.isCloudMode() || !backend.isDecommissioning())
&& (!Config.isCloudMode() || cluster == null
|| cluster.equals(backend.getCloudClusterName()))) {
if (isBackendAvailable(backend, cluster)) {
return backend.getId();
} else {
tableToBeMap.remove(encode(cluster, tableId));
Expand All @@ -372,13 +369,26 @@ private Long getCachedBackend(String cluster, long tableId) {
return null;
}

private boolean isBackendAvailable(Backend backend, String cluster) {
if (backend == null || !backend.isAlive() || backend.isDecommissioned() || !backend.isLoadAvailable()) {
return false;
}
if (!Config.isCloudMode()) {
return true;
}
Copy link
Contributor

@github-actions github-actions bot Mar 20, 2026

Choose a reason for hiding this comment

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

[Observability issue] Now that isBackendAvailable() rejects backends based on isLoadAvailable() (which checks isLoadDisabled and isShutDown), the error messages in selectBackendForCloudGroupCommitInternal (line ~303) and selectBackendForLocalGroupCommitInternal (line ~338) should include isLoadDisabled and isShutDown fields. Currently they only log alive, active, decommissioned (and decommissioning in the cloud version).

Without this, when all backends are alive but load-disabled, operators will see a misleading error showing all backends as healthy.

Suggested addition to the error message map lambdas:

+ ", loadDisabled=" + be.isLoadDisabled() + ", shutdown=" + be.isShutDown()

(Note: isShutDown() is private — you may need to add isLoadAvailable() to the message instead, or make isShutDown accessible.)

// for cloud mode
if (backend.isDecommissioning()) {
return false;
}
return cluster == null || cluster.equals(backend.getCloudClusterName());
}

@Nullable
private Long getRandomBackend(String cluster, long tableId, List<Backend> backends) {
OlapTable table = (OlapTable) Env.getCurrentEnv().getInternalCatalog().getTableByTableId(tableId);
Collections.shuffle(backends);
for (Backend backend : backends) {
if (backend.isAlive() && !backend.isDecommissioned() && (!Config.isCloudMode()
|| !backend.isDecommissioning())) {
if (isBackendAvailable(backend, cluster)) {
tableToBeMap.put(encode(cluster, tableId), backend.getId());
tableToPressureMap.put(tableId,
new SlidingWindowCounter(table.getGroupCommitIntervalMs() / 1000 + 1));
Expand Down
Loading