Control Mnesia database size

[rkdm91]

Hello,

I would like to put the database in Ram, for that i need to control the size of Mnesia database. I suppose that the database size depends on system parameters ( number of device, gateway, profile ,etc) which can be controlled, and it depends also on uplink/downlink and event saved.

As my uplinks and events are transferred to my own server, i don't need them to be store on the lorawan server , i would like to limit the number of uplink/downlink and event saved in the database (the fact that i can miss some uplink/event is not a problem for me), do you think it is possible ? I know that we can configure :

% amount of rxframes retained for each device/node
{retained_rxframes, 50},

form the sys.config file but more than 50 uplinks are available on my lora-server, I don't understand the use of this parameter.

Thank you in advance

[gotthardp]

Hello. You should have no more than 50 frames for each device. If you have 100 different devices, you will have no more than 5000 stored frames. You should be able to decrease this to 0, but it will impact the timeline and RX quality graphs for your devices. (You will not see anything there.)

[rkdm91]

Hello, I confirmed that I have only one device commissioned and i have more than 50 uplinks frame saved :

[gotthardp]

and did you change the retained_rxframes settings or not?

[rkdm91]

I didn't change the parameter, if i am correct the default setting is :

% amount of rxframes retained for each device/node
{retained_rxframes, 50}

which is the case in my /lorawan-server/releases/0.6.2/sys.config file

[hallard]

I confirm the issue,

My config is as follow

    % amount of rxframes retained for each device/node
    {retained_rxframes, 10},
    % events duration on interface
    {event_lifetime, 3600},

Looks like got 148 frames (from same device), since got plenty of devices on multitech GW it really slow down server. Of course service and GW has been restarted

[gotthardp]

The trimming works for me. Please note that:

• the server stores 50 uplinks and 50 downlinks, so 100 in total (this matches the @rkdm91 screenshot)
• the trimming is done every hour, so if the traffic is high I can imagine there will be 100 frames left during last purging, plus 48 frames received since that

@hallard, could you please double check you really have more than 100 frames that older than 1 hour?

[hallard]

@gotthardp My setup is 10 frames retained not 100, did I missed something?

[gotthardp]

@hallard How old the frames are and for how long was the server running?

[hallard]

server is rebooted every day, also I purged all so can't answer.

anyway, recompiled installed, and setup your debug config, all seems good on config, need to check after time

mtcdt:~$ cat /var/log/lorawan-server/debug.log | grep Using
2019-03-27 17:35:13.238 [debug] <0.313.0>@lorawan_app:start:17 Using config: [{retained_rxframes,10},{http_admin_redirect_ssl,true},{gateway_delay,200},{deduplication_delay,200},{trim_interval,1800},{event_lifetime,1800},{packet_forwarder_listen,[{port,1680}]},{ssl_options,[]},{websocket_timeout,infinity},{http_custom_web,[{"/",file,<<"/home/mtadm/html/index.html">>,[{<<"anonymous">>,'*'}]},{"/[...]",dir,<<"/home/mtadm/html">>,[{<<"anonymous">>,'*'}]}]},{devstat_gap,{432000,96}},{http_admin_credentials,{<<"admin">>,<<"admin">>}},{slack_server,{"slack.com",443}},{http_extra_headers,#{}},{connectors,[{lorawan_connector_amqp,[<<"amqp">>,<<"amqps">>]},{lorawan_connector_mqtt,[<<"mqtt">>,<<"mqtts">>]},{lorawan_connector_http,[<<"http">>,<<"https">>]},{lorawan_connector_mongodb,[<<"mongodb">>]},{lorawan_connector_ws,[<<"ws">>]}]},{frames_before_adr,50},{http_admin_listen_ssl,[{port,8443},{certfile,"cert.pem"},{cacertfile,"cacert.pem"},{keyfile,"key.pem"}]},{http_admin_path,<<"/admin">>},{max_lost_after_reset,10},{applications,[{<<"semtech-mote">>,lorawan_application_semtech_mote}]},{http_admin_listen,[{port,80}]},{server_stats_interval,60}]