Broken queues after crash

[Luke Gee]

Hi,

After what looks like a crash after a server restart we have a number of queues that are unresponsive to the point I cant even delete them via the management API. They are showing no statuses and all stats have '?' when viewing the queue in the management console. Having a look around it looks very similar to the issue stated here: https://groups.google.com/g/rabbitmq-users/c/dgjikJDu_ZM

Luckily this is in a test environment but would be keen to try and understand the issue so that I can ensure it doesnt happen in prod.

Our setup is a 2 node cluster,  running on prem on windons VMs. Running version 3.8.0 on erl 22.1

Ive attached what seems to be the relevant logs and any help would be much appreciated. TIA

[John Pfersich]

well, for one thing odd numbers of RMQ servers are the recommended configuration. and RMQ is currently at version 3.8.12, have you considered upgrading?

/—————————————————————/

For encrypted mail use jgpfe...@protonmail.com - Free account at ProtonMail.com

Web: https://objectnets.net and https://objectnets.org

https://datascilv.com https://datascilv.org

On Feb 12, 2021, at 00:33, Luke Gee <luke....@gmail.com> wrote:

Hi,

--
You received this message because you are subscribed to the Google Groups "rabbitmq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rabbitmq-user...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/rabbitmq-users/65c7b671-403b-479d-bb67-f4bbaf6960fen%40googlegroups.com.

<rabbitmq.log>

[Diana Corbacho]

Hi Luke,

You're probably hitting https://github.com/rabbitmq/rabbitmq-server/pull/2273, so please upgrade to the latest patch release (currently 3.8.11).

As mentioned above, it is worth reviewing your cluster size [1]. Also I would consider if the usage of the new quorum queue type [2] might be better for your use case.

[1] https://www.rabbitmq.com/clustering.html

[2] https://www.rabbitmq.com/quorum-queues.html