Disk is almost full and snapshot_index is still -1

[Vishnu Vardhan]

Hi Team,

I am using RabbitMQ 4.0.5 with quorum queues in the system. I am seeing that my disk space is almost full and currently there are no messages in the current quorum queues as they all are consumed, also the count of unacknowledged message is 0.

I was going through the previous issues raised on similar issue, and found some information on "rabbitmq-diagnostics quorum_status <queue_name>" command and the fields "commit_index" "snapshot_index". 

commit_index  >> snapshot_index indicates Compaction is lagging.

But in our case the snapshot_index is -1 and the information I found is that compaction is not yet started.

So what exactly is stopping the compaction of these segments, can it be triggered manually in anyway so that  the disk is not fully filled.

Also are there any guidelines on the disk space when quorum queues are used, I found this doc link https://www.rabbitmq.com/docs/quorum-queues#disk-space, but it does not provide a performance benchmark for disk size.

Please provide more information on the above queries. Thanks in advance.

[Vishnu Vardhan]

Adding more info here

there are 2 segment files of large sizes and that the followers have caught up to the leaders (observations from cluster status)

So pls suggest on why has compaction not occurred, what forces rmq to compact, how does a user force rmq to trigger a compaction. what are the ways to work around this issue and how to mitigate such issues