Maximum heap size reached

[badsector3]

Hi.

When I send a message of size 160KB (2 per second) the connection is dropped with following error:

2024-02-20T12:05:46.242709+00:00 [error] Process:            <0.14133.0> on node 'emqx@emqx-cluster-core-xxx-1.emqx-cluster-headless.example.svc.cluster.local', Context:            maximum heap size reached, Max Heap Size:      6291456, Total Heap Size:    7576612, Kill:               true, Error Logger:       true, Message Queue Len:  0, GC Info:            [{old_heap_block_size,2984878},{heap_block_size,4560232},{mbuf_size,80000},{recent_size,1944329},{stack_size,63},{old_heap_size,0},{heap_size,2024335},{bin_vheap_size,29161},{bin_vheap_block_size,46422},{bin_old_vheap_size,0},{bin_old_vheap_block_size,46422}]

If I disable retained messages then there is no issue.
Config:

delayed {enable = true, max_delayed_messages = 1000}
retainer {
  backend {
    index_specs = [
      [1, 2, 3],
      [1, 3],
      [2, 3],
      [3]
    ]
    max_retained_messages = 1000
    storage_type = disc
    type = built_in_database
  }
  enable = true
  max_payload_size = 1MB
  msg_clear_interval = 0s
  msg_expiry_interval = 0s
  stop_publish_clear_msg = false
}

I tried with:

    max_retained_messages = 0
    storage_type = ram 

No change. Any ideas?

[badsector3]

Update to the above - disable retain messages doesn't change anything. Still the same after some additional tests

[terry-xiaoyu]

Have you enabled the debug level log message or have you enabled log tracing for that clientid? Printing details for a client will consume lots of memory.

[badsector3]

Thanks @terry-xiaoyu It was exactly that - debug logs.
Found as well ticket on EMQX about it emqx/emqx#9867 (comment)

Problem solved - disable the debug logs in my case fixed the issue.