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.

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

fix potential ArgumentException during shard rebalancing #7367

Merged
merged 2 commits into from
Oct 28, 2024
Merged

fix potential ArgumentException during shard rebalancing #7367

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0d8bc07
fix potential `ArgumentException` during shard rebalancing
Aaronontheweb Oct 28, 2024
821e4bf
moved back to original discard design
Aaronontheweb Oct 28, 2024
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
8 changes: 7 additions & 1 deletion src/contrib/cluster/Akka.Cluster.Sharding/Internal/AbstractLeastShardAllocationStrategy.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -126,7 +126,13 @@ protected bool IsAGoodTimeToRebalance(IEnumerable<RegionEntry> regionEntries)

protected ImmutableList<RegionEntry> RegionEntriesFor(IImmutableDictionary<IActorRef, IImmutableList<string>> currentShardAllocations)
{
var addressToMember = ClusterState.Members.ToImmutableDictionary(m => m.Address, m => m);
Copy link
Member Author

Choose a reason for hiding this comment

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

Left some comments that explain the change

// switched to using `GroupBy` instead just ToImmutableDictionary due to https://github.com/akkadotnet/akka.net/issues/7365
// it's very rare, but possible, that there can be two members with the same address in the ClusterState. This can happen
// when a node quickly reboots and re-uses its old address, but the old incarnation hasn't been downed yet.
var addressToMember = ClusterState.Members
.GroupBy(m => m.Address)
// using Last or First here is non-deterministic since the UID that appears in the UniqueAddress sort order is random
.ToImmutableDictionary(g => g.Key, g => g.First());

return currentShardAllocations.Select(i =>
{
Expand Down
4 changes: 2 additions & 2 deletions src/contrib/cluster/Akka.Cluster.Sharding/ShardAllocationStrategy.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -169,11 +169,11 @@ public override Task<IImmutableSet<ShardId>> Rebalance(IImmutableDictionary<IAct
var sortedRegionEntries = RegionEntriesFor(currentShardAllocations).OrderBy(i => i, ShardSuitabilityOrdering.Instance).ToImmutableList();
if (IsAGoodTimeToRebalance(sortedRegionEntries))
{
var (_, Shards) = MostSuitableRegion(sortedRegionEntries);
var (_, shards) = MostSuitableRegion(sortedRegionEntries);
// even if it is to another new node.
var mostShards = sortedRegionEntries.Select(r => r.ShardIds.Where(s => !rebalanceInProgress.Contains(s))).MaxBy(i => i.Count())?.ToArray() ?? Array.Empty<string>();

var difference = mostShards.Length - Shards.Count;
var difference = mostShards.Length - shards.Count;
if (difference >= _rebalanceThreshold)
{
var n = Math.Min(
Expand Down