Skip to content

Coordinator fix balancer stuck #5987

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

Merged
jon-wei merged 6 commits into from
Jul 12, 2018
Merged

Coordinator fix balancer stuck #5987

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
6dd4533
this will fix it
clintropolis Jul 9, 2018
0253fd7
filter destinations to not consider servers already serving segment
clintropolis Jul 9, 2018
b3bcfde
fix it
clintropolis Jul 10, 2018
a0c2714
cleanup
clintropolis Jul 10, 2018
8e84ed0
fix opposite day in ImmutableDruidServer.equals
clintropolis Jul 10, 2018
7b509d2
simplify
clintropolis Jul 10, 2018
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
6 changes: 1 addition & 5 deletions server/src/main/java/io/druid/client/ImmutableDruidServer.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -142,11 +142,7 @@ public boolean equals(Object o)

ImmutableDruidServer that = (ImmutableDruidServer) o;

if (metadata.equals(that.metadata)) {
return false;
}

return true;
return metadata.equals(that.metadata);
}

@Override
Expand Down
56 changes: 42 additions & 14 deletions server/src/main/java/io/druid/server/coordinator/helper/DruidCoordinatorBalancer.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -93,9 +93,11 @@ private void balanceTier(
CoordinatorStats stats
)
{
final BalancerStrategy strategy = params.getBalancerStrategy();
final int maxSegmentsToMove = params.getCoordinatorDynamicConfig().getMaxSegmentsToMove();

if (params.getAvailableSegments().size() == 0) {
log.info("Metadata segments are not available. Cannot balance.");
return;
}
currentlyMovingSegments.computeIfAbsent(tier, t -> new ConcurrentHashMap<>());

if (!currentlyMovingSegments.get(tier).isEmpty()) {
Expand All @@ -117,33 +119,59 @@ private void balanceTier(
numSegments += sourceHolder.getServer().getSegments().size();
}


if (numSegments == 0) {
log.info("No segments found. Cannot balance.");
return;
}

final BalancerStrategy strategy = params.getBalancerStrategy();
final int maxSegmentsToMove = Math.min(params.getCoordinatorDynamicConfig().getMaxSegmentsToMove(), numSegments);
final int maxIterations = 2 * maxSegmentsToMove;
final int maxToLoad = params.getCoordinatorDynamicConfig().getMaxSegmentsInNodeLoadingQueue();
long unmoved = 0L;
for (int moved = 0; (moved + unmoved) < maxSegmentsToMove;) {
final BalancerSegmentHolder segmentToMove = strategy.pickSegmentToMove(toMoveFrom);

if (segmentToMove != null && params.getAvailableSegments().contains(segmentToMove.getSegment())) {
final List<ServerHolder> toMoveToWithLoadQueueCapacity =
for (int moved = 0, iter = 0; (moved + unmoved) < maxSegmentsToMove; ++iter) {
final BalancerSegmentHolder segmentToMoveHolder = strategy.pickSegmentToMove(toMoveFrom);

if (segmentToMoveHolder != null && params.getAvailableSegments().contains(segmentToMoveHolder.getSegment())) {
final DataSegment segmentToMove = segmentToMoveHolder.getSegment();
final ImmutableDruidServer fromServer = segmentToMoveHolder.getFromServer();
// we want to leave the server the segment is currently on in the list...
// but filter out replicas that are already serving the segment, and servers with a full load queue
final List<ServerHolder> toMoveToWithLoadQueueCapacityAndNotServingSegment =
toMoveTo.stream()
.filter(s -> maxToLoad <= 0 || s.getNumberOfSegmentsInQueue() < maxToLoad)
.filter(s -> s.getServer().equals(fromServer) ||
(!s.isServingSegment(segmentToMove) &&
(maxToLoad <= 0 || s.getNumberOfSegmentsInQueue() < maxToLoad)))
.collect(Collectors.toList());

final ServerHolder destinationHolder =
strategy.findNewSegmentHomeBalancer(segmentToMove.getSegment(), toMoveToWithLoadQueueCapacity);

if (destinationHolder != null) {
moveSegment(segmentToMove, destinationHolder.getServer(), params);
moved++;
if (toMoveToWithLoadQueueCapacityAndNotServingSegment.size() > 0) {
final ServerHolder destinationHolder =
strategy.findNewSegmentHomeBalancer(segmentToMove, toMoveToWithLoadQueueCapacityAndNotServingSegment);

if (destinationHolder != null && !destinationHolder.getServer().equals(fromServer)) {
moveSegment(segmentToMoveHolder, destinationHolder.getServer(), params);
moved++;
} else {
log.info("Segment [%s] is 'optimally' placed.", segmentToMove.getIdentifier());
unmoved++;
}
} else {
log.info("Segment [%s] is 'optimally' placed.", segmentToMove.getSegment().getIdentifier());
log.info(
"No valid movement destinations for segment [%s].",
segmentToMove.getIdentifier()
);
unmoved++;
}
}
if (iter >= maxIterations) {
log.info(
"Unable to select %d remaining candidate segments out of %d total to balance after %d iterations, ending run.",
(maxSegmentsToMove - moved - unmoved), maxSegmentsToMove, iter
);
break;
}
}

if (unmoved == maxSegmentsToMove) {
Expand Down