3.13 with classic mirrored queues runs into an einval in rabbit_msg_store:reader_close/1

[ko2024]

Describe the bug

I have a scenario in my cluster where it lost queue replication, several queues were left with "?", problem sync, and it was only possible to reestablish the cluster by terminating the asg nodes.

My version is 3.13.7 and the error I got in the log indicates a disk problem. Is there a related bug in this version?

Reproduction steps

1.I can't reproduce
2.
3.
...

Expected behavior

** Reason for termination ==
** {{badmatch,{error,einval}},
[{rabbit_msg_store,reader_close,1,
[{file,"rabbit_msg_store.erl"},{line,708}]},
{rabbit_variable_queue,delete_and_terminate,2,
[{file,"rabbit_variable_queue.erl"},{line,544}]},
{rabbit_priority_queue,delete_and_terminate,2,
[{file,"rabbit_priority_queue.erl"},{line,185}]},
{gen_server2,terminate,3,[{file,"gen_server2.erl"},{line,1158}]},
{proc_lib,wake_up,3,[{file,"proc_lib.erl"},{line,251}]}]}
** In 'terminate' callback with reason ==
** {{badmatch,

0.254349487.1> ** In 'terminate' callback with reason ==7.861386+00:00 [error] <0.254349487.1> ** {{badmatch,

generated several errors like the example below:

rabbit_msg_store_ets_index, 0 [error] <0.254349487.1> <<"/var/lib/rabbitmq/mnesia/node/msg_stores/vhosts/628WB79CIFDYO9LJI6DKMI09L/msg_store_persistent">>,

Additional context

No response

[kjnilsson]

Only RabbitMQ version 4.1 is available for community support. Please upgrade to this version (which will mean you need to migrate away from classic mirrored queues) and see if the problem remains.

https://www.rabbitmq.com/release-information

Cheers
Karl