NPE after queue draining

[1ou]

We are using a custom circuit breaker mechanism over lettuce, when state is changed from CLOSED to OPEN we call this method for queue draining. The circuit breaker stays in OPEN state for 30 seconds, so new commands are not added at this time period.
The chosen reliability strategy is: Reliability.AT_LEAST_ONCE

commandHandler.getEndpoint().notifyDrainQueuedCommands(commandHandler);

CommandHandler and DefaultEndpoints contains in queues\stack only MGET commands, such error occurs after queue draining:

Caused by: java.util.NoSuchElementException
	at java.base/java.util.ArrayList$Itr.next(ArrayList.java:970)
	at io.lettuce.core.output.KeyValueListOutput.set(KeyValueListOutput.java:80)
	at io.lettuce.core.protocol.RedisStateMachine.safeSet(RedisStateMachine.java:810)
	at io.lettuce.core.protocol.RedisStateMachine.handleBytes(RedisStateMachine.java:572)
	at io.lettuce.core.protocol.RedisStateMachine$State$Type.handle(RedisStateMachine.java:206)
	at io.lettuce.core.protocol.RedisStateMachine.doDecode(RedisStateMachine.java:334)
	at io.lettuce.core.protocol.RedisStateMachine.decode(RedisStateMachine.java:295)
	at io.lettuce.core.protocol.CommandHandler.decode(CommandHandler.java:841)
	at io.lettuce.core.protocol.CommandHandler.decode0(CommandHandler.java:792)
	at io.lettuce.core.protocol.CommandHandler.decode(CommandHandler.java:775)
	at io.lettuce.core.protocol.CommandHandler.decode(CommandHandler.java:658)
	at io.lettuce.core.protocol.CommandHandler.channelRead(CommandHandler.java:598)
	at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:379)
	at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:365)
	at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:357)
	at io.netty.channel.DefaultChannelPipeline$HeadContext.channelRead(DefaultChannelPipeline.java:1410)
	at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:379)
	at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:365)
	at io.netty.channel.DefaultChannelPipeline.fireChannelRead(DefaultChannelPipeline.java:919)
	at io.netty.channel.epoll.AbstractEpollStreamChannel$EpollStreamUnsafe.epollInReady(AbstractEpollStreamChannel.java:795)
	at io.netty.channel.epoll.EpollEventLoop.processReady(EpollEventLoop.java:480)
	at io.netty.channel.epoll.EpollEventLoop.run(EpollEventLoop.java:378)
	at io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:986)
	at io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74)
	at io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30)

Lettuce version is: 6.1.2.RELEASE

My assumption is that Iterable has been cleared from heap due to queues draining, but Redis commands hadn't been cleared yet somehow.
Could you help me with this issue? Not sure that I'm draining queues correctly, maybe there is a more correct solution.
By the way it is only occurred when there is list of keys, I've tested SETEX command, and it is doing well.

[mp911de]

Without a reproducer it's difficult to find out what is happening here. Either the protocol is out of sync or the command doesn't match the Redis response.

My assumption is that Iterable has been cleared from heap due to queues draining

Endpoint.notifyDrainQueuedCommands(…) is being called only as consequence of a disconnect (ChannelInboundHandler.channelInactive). It mustn't be called externally as there's no synchronization and the lack of synchronization may cause problematic states.

[1ou]

Thank you for reply!
We have a scenario at high load when stack of commands at CommandHandler is full, as a result circuit breaker is open (it is triggered then stack is full by our logic), because we chose Reliability.AT_LEAST_ONCE strategy for auto reconnect supporting, once circuit breaker is close again all unsent commands from disconnectBuffer and CommandHandler.stack are going to be sent again, but they are outdated we don't need them.
What is the proper way of handling such scenario, how can we get rid of all those commands?