François Travais

the `ps` shows one process running but the node is not responding and not usable. And I don't now much about the inner mechanisms of erlang and VerneMQ but the...

@ioolkos The same happened again last night: ``` $ systemctl status vernemq ● vernemq.service - VerneMQ Server Loaded: loaded (/etc/systemd/system/vernemq.service; enabled; vendor preset: disabled) Active: active (running) since Tue 2020-05-19...

The proof that something is wrong is that `systemctl stop vernemq` fail in this state: ``` $ systemctl stop vernemq $ systemctl status vernemq ● vernemq.service - VerneMQ Server Loaded:...

This message is shown systematically, even when everything is fine. I guess there's something wrong with the systemd service file definition of vernemq

@ioolkos I'm using the service file embedded in the RPM you provide. ``` # systemd unit file [Unit] Description=VerneMQ Server After=network.target [email protected] Wants=network.target [email protected] [Service] Type=forking PIDFile=/run/vernemq/vernemq.pid User=vernemq Group=vernemq NotifyAccess=all...

Here are the queues messages, the green line is the node 1, which died.  Here's the VM memory. 

Here's another example:  

I don't have the amount of messages stored, because this information isn't available on the prometheus metrics (see #1488)

The "queued messages" graph uses this formula: `queue_message_in - queue_message_out - queue_message_expired - queue_message_unhandled`

There has been a change this time: even if the VerneMQ died, restarting it worked. For the two examples above VerneMQ is now up. In the past it, the same...