Queue has crashed and failed to restart

[MakarMS]

Describe the bug

2025-04-03 06:44:09.358376+00:00 [error] <0.516794.0> Channel error on connection <0.516786.0> (xxx.xxx.x.xx:49522 -> xxx.xxx.x.xx:5672, vhost: 'xx', user: 'xxx'), channel 1:
2025-04-03 06:44:09.358376+00:00 [error] <0.516794.0> operation basic.get caused a channel exception not_found: queue 'test_queue' in vhost 'xx' has crashed and failed to restart

I have a monitor running that tries every minute to send a message and then read it. The problem is that sometimes these errors appear. After crashing and failing to restart, the queue becomes writable and readable again, with the parameters that were defined in the schema files. I.e. the queue is somehow restored and continues working. The problem appeared after upgrading from 3.11 to 4.0.4. The upgrade was done by uninstalling rabbitmq and installing the latest available version from the package manager.

Reproduction steps

1. Creating a queue
2. Periodic writes and reads
3. Waiting for random error

Expected behavior

It is expected that no random write and read errors will occur

Additional context

OS: FreeBSD
Uptime (seconds): 592033
Is under maintenance?: false
RabbitMQ version: 4.0.4
RabbitMQ release series support status: see https://www.rabbitmq.com/release-information
Erlang configuration: Erlang/OTP 27 [erts-15.1.2] [source] [64-bit] [smp:40:40] [ds:40:40:10] [async-threads:1] [jit:ns] [dtrace] [sharing-preserving]
Crypto library: OpenSSL 3.0.16 11 Feb 2025
Erlang processes: 411 used, 1048576 limit
Scheduler run queue: 1
Cluster heartbeat timeout (net_ticktime): 60

{
      "arguments": {
        "x-queue-type": "classic",
        "x-message-ttl": 1000
      },
      "auto_delete": false,
      "durable": true,
      "name": "test_queue",
      "type": "classic",
      "vhost": "xx"
}

[lukebakken]

The problem appeared after upgrading from 3.11 to 4.0.4. The upgrade was done by uninstalling rabbitmq and installing the latest available version from the package manager

Upgrading 3.11 to 4.0.4 directly is NOT supported. You MUST upgrade to each intermediate version first, and use the latest patch version for each intermediate - https://www.rabbitmq.com/docs/upgrade

[mkuratczyk]

There's almost certainly more in the logs - a stack trace, which would tell us much more than anything else can.

[MakarMS]

Unfortunately, there was no other information in the logs other than the two error records I provided in the description

[MakarMS]

It looks weird

[mkuratczyk]

There was one queue crash bugfix in 4.0.6, so step 1 would be to upgrade to the latest version and see if it still happens.

[MakarMS]

I'll try updating the cluster tomorrow morning and see, thanks

[MakarMS]

Hey @mkuratczyk. Updated the cluster to 4.0.6, let's wait a couple of days to see if the error shows up or not

[MakarMS]

Hi all, the error didn't happen again. Apparently it was just this bug.

[MakarMS]

Thank you all