fix: Improve cluster connection pool logic when disconnecting#5

Merged

mcollina merged 20 commits intovalkey-io:mainfrom martinslota:clean-up-node-listeners-upon-disconnect

Jun 13, 2024

+129-71

Commits on Jun 10, 2024

Tell Redis cluster to disable protected mode before running tests
martinslota
committed
Try to enable Redis cluster tests on CI
martinslota
committed
Add a failing test around Redis cluster disconnection logic
martinslota
committed
Rename function parameter
martinslota
committed
Turn node error listener into an arrow function so that points to the connection pool instance
martinslota
committed
Extract node listeners into separate constants
martinslota
committed
Keep track of listeners along with each Redis client
martinslota
committed
Remove node listeners when the node is being removed
martinslota
committed
Emit node removal events whenever a node is removed
martinslota
committed
When resetting, add nodes before removing old ones
martinslota
committed
Rename Node type to NodeRecord for clarity
martinslota
committed
Also rename the field holding node records
martinslota
committed
Rename variable to nodeRecord
martinslota
committed
Rename another variable to nodeRecord
martinslota
committed
Fix a reference to connection pool nodes
martinslota
committed
Do not fail when retrieving a node by non-existing key
martinslota
committed
Revert "Fix a reference to connection pool nodes"
martinslota
committed
Fix a reference to connection pool nodes, this time a bit more correctly
martinslota
committed
Add a valid slots table to mock server in tests that expect to be able to connect using the Cluster client
martinslota
committed
Do not assume that node removal will occur *after* refreshSlotsCache() is finished
martinslota
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: Improve cluster connection pool logic when disconnecting#5

Commits on Jun 10, 2024

Tell Redis cluster to disable protected mode before running tests

Try to enable Redis cluster tests on CI

Add a failing test around Redis cluster disconnection logic

Rename function parameter

Turn node error listener into an arrow function so that points to the connection pool instance

Extract node listeners into separate constants

Keep track of listeners along with each Redis client

Remove node listeners when the node is being removed

Emit node removal events whenever a node is removed

When resetting, add nodes before removing old ones

Rename Node type to NodeRecord for clarity

Also rename the field holding node records

Rename variable to nodeRecord

Rename another variable to nodeRecord

Fix a reference to connection pool nodes

Do not fail when retrieving a node by non-existing key

Revert "Fix a reference to connection pool nodes"

Fix a reference to connection pool nodes, this time a bit more correctly

Add a valid slots table to mock server in tests that expect to be able to connect using the Cluster client

Do not assume that node removal will occur after refreshSlotsCache() is finished