For another option, the strategy I took was that I created global scripts to create the data points (MQTT/Modbus, virtual, meta) for my devices.

The scripts I have are:
dataPointJsonTempates
--> This script contains the minimal set of values I need to set for each type of data point.
createDataPointFunction
--> This script contains the list of points I need for each of the devices. It loops through each list of points and calls the appropriate function below.
createDeviceFunctions
--> This script creates each of the appropriate data point, event detector, and event handlers.

The intention was to detect the error generated when an unknown MQTT topic is subscribed to and create the required data points from the device id in the topic. I haven't gotten that working yet but it is a lower priority item at the moment.

For now, I have a scripting data source to call the function:
createDevice('PCU_110006');

I have created multiple 'createDataPointFunctions' scripts for the different devices I need. The same pattern is applied.

Ian

Hi @dgm

You are close:

            options="{
                synchronizeGrid: false, 
                valueAxes:[
                    {title: 'Connected', titleColor: '#ff4d4d', minimum: 0, maximum: 1, integersOnly: true, baseValue: 2},
                    {title: '°C', titleColor: '#ff4d4d', minimum: -50, maximum: 50}
                ]
            }"

Each axis is an element in the valueAxes array.

Ian

Heya,

v3.7.7

For some time, I was encountering an odd behavior in displaying the value of data points for a device. Some devices would be fine and some would have missing data points. I confirmed that the data points I expected were in the data source and had a valid value but for some reason were not being returned in the point query.

Finally I had some time to devote to investigate this issue and discovered that there is an implicit limit of 100 on a data point query.

For example, the number of data points returned by the follow queries should be the same:

<ma-point-query query="{deviceName: $ctrl.deviceName}" sort="'name'" points="ptstmp"></ma-point-query>
*{{ptstmp.length}}*
<ma-point-query query="{deviceName: $ctrl.deviceName}" sort="'name'" limit="300" points="ptsDevice"></ma-point-query>
*{{ptsDevice.length}}*

However, the lengths of the arrays where 100 and 153 respectively.

According to the API docs in the application:
"All data points added to your Mango system will display by default, unless you set the limit property."
and
"Limit takes place after query and sorting (no limit by defualt)"

The docs would indicate that if no limit is specified, then all of the data points matching the query should be returned.

I am not sure if there is a possible issue with the docs or the behavior of the point query function.

The issue can be fixed by adding limit="-1" to the query but it is a maintenance challenge to update all uses of the point query in the app.

Thanks
Ian

Heya,

What is the best method for disabling or removing fonts that are unneeded, such as greek, latin, vietnamese, etc? Loading these fonts can take some time over a cellular modem and makes the initial connection to the application appear slow.

I found 2 packages, OpenSans and Roboto, under /web/resources/fonts but I am unsure the best way to disable them. I am a little hesitant to remove them all together as I don't know their impact on the rest of the application.

Any suggestions?

Thanks
Ian

Heya,

I seem to be struggling with the syntax and layout of a model for use with Freemarker.

The list of data points returned from DataPointQuery works as-is but if I try to create my own model, based on the data points list, I get the error: The value you try to list is an extended_hash (wrapper: f.t.SimpleHash), thus you must specify two loop variables after the "as"; one for the key, and another for the value, like <#... as k, v>).

I want to create a model as I want to include information that isn't available in the records for each data point returned in the query.

To me, this simple model appears similar to the data points but does not work:

var tmpRecs = [
    {deviceName: 'PCU_110020', path: 'PCU_110090'},
    {deviceName: 'PCU_110021', path: 'PCU_110090'},
    {deviceName: 'PCU_110022', path: 'PCU_110090'}
];

Most of the information I have found on the net is more Java related than specifically javascript.

Does anyone have an example of a model in javascript they have built or some pointers as to what I am doing wrong? heh

Any suggestions would be appreciated.

Thanks
Ian

Heya,

Is there a setting for the number of retries that will be taken before an email fails? It appears that only 1 attempt to send the email is made, unless the app is restarted.

Hopefully its just something I have missed.

Thanks
Ian

Hi @MattFox

Good suggestion to attempt to manually import the changes to the data point. When I did that, I got the "errors":

Data point 'DP_xyz': intervalLoggingPeriod --> Must be greater than zero
Data point 'DP_xyz': purgePeriod --> Must be greater than zero
Data point 'DP_xyz': intervalLoggingSampleWindowSize --> Must be greater than zero
Import complete

It's odd that the import tool seems to imply that the import was successful. /shrug When I examine the fields, those values are set to zero, yet the data point was originally saved. Hmmm

The next test was I exported the point from the ui instead and imported. Same "errors".

Next, I just tried to save the data point in the ui and this time got an error! Hmmm

What I discovered is that the copied data point I was using for testing didn't actually save properly. It looked like it had. The issue was that the copy process didn't copy the custom template used in the original. When I re-copied the point, the template wasn't applied and caused an error when saving. After selecting the template, the save was "successful".

Then my function worked just fine to update the name. A bit of an adventure figuring that one out heh.

Thanks
Ian

Heya,

I am trying to create a function that will update the name of a data point. I have successfully used a similar function to create a point but this one does not appear to work:

function updateDataPointName(deviceName, pointName, newName) {

    var dataPoints = JSON.parse(JsonEmport.getConfiguration("dataPoints")).dataPoints;
    // LOG.info(JSON.stringify(dataPoints));
    
    function findByName(list, deviceName, pointName) {
        for (var k = 0; k < list.length; k+=1)
            if (list[k].deviceName === deviceName && list[k].name === pointName)
                return list[k];
        return undefined;
    }

    var point = findByName(dataPoints, deviceName, pointName);
    if (typeof point === 'undefined') {
        LOG.info('Could not find point with name '+deviceName+' - '+pointName);
    }
    else {
        point.name = newName;
        LOG.info(JSON.stringify(point));
    
        JsonEmport.setImportDuringValidation(true);
        JsonEmport.doImport(JSON.stringify({"dataPoints": [point]}));
    }
}

The logs show that the point name is being changed but when I check the point in the data source, the name was not changed. I don't see any errors in the ma.log either.

Any suggestions? Can you change the data point name this way?

Thanks
Ian

Heya,

(v3.6.5)

I have been tracking an issue that causes the MQTT data source to disconnect from its broker.

What I have been seeing in the ma.log is:

WARN  2020-02-21T08:38:12,594 (com.serotonin.m2m2.rt.dataSource.PollingDataSource.incrementUnsuccessfulPolls:150) - Data Source Application Virtual aborted 3093 polls since it started. 
WARN  2020-02-21T08:45:15,170 (com.serotonin.m2m2.rt.dataSource.PollingDataSource.incrementUnsuccessfulPolls:150) - Data Source Mango Internal aborted 1494 polls since it started. 
WARN  2020-02-21T08:45:15,191 (com.serotonin.m2m2.rt.dataSource.PollingDataSource.incrementUnsuccessfulPolls:150) - Data Source Internal Stats aborted 22 polls since it started. 
ERROR 2020-02-21T08:45:15,249 (com.infiniteautomation.m2m2.mqtt.client.MqttClientDataSourceRT$1.connectionLost:632) - MQTT Connection Lost 
org.eclipse.paho.client.mqttv3.MqttException: Connection lost
	at org.eclipse.paho.client.mqttv3.internal.CommsSender.handleRunException(CommsSender.java:153) ~[?:?]
	at org.eclipse.paho.client.mqttv3.internal.CommsSender.run(CommsSender.java:138) ~[?:?]
	at java.lang.Thread.run(Thread.java:748) [?:1.8.0_212]
Caused by: java.net.SocketException: Broken pipe (Write failed)
	at java.net.SocketOutputStream.socketWrite0(Native Method) ~[?:1.8.0_212]
	at java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:111) ~[?:1.8.0_212]
	at java.net.SocketOutputStream.write(SocketOutputStream.java:155) ~[?:1.8.0_212]
	at java.io.BufferedOutputStream.flushBuffer(BufferedOutputStream.java:82) ~[?:1.8.0_212]
	at java.io.BufferedOutputStream.flush(BufferedOutputStream.java:140) ~[?:1.8.0_212]
	at org.eclipse.paho.client.mqttv3.internal.wire.MqttOutputStream.flush(MqttOutputStream.java:49) ~[?:?]
	at org.eclipse.paho.client.mqttv3.internal.CommsSender.run(CommsSender.java:107) ~[?:?]
	... 1 more

These errors seemed to occur on a fairly regular timing of about 3.5 hours. It wasn't until yesterday when I got the Mango internal source working that I saw that it was very regular.

The log entries occur when the free memory is "reset" back to 1.2gb. The one anomaly in the middle is when the backup/purge occurs, at which time there are log entries regarding the aborted polling the mqtt lost connection.

I have 2 environment setup with a very similar load:
VM's with 2GB memory allocated, 4GB swap
JVM: ~80% memory usage, 0-15% CPU usage
JAVAOPTS="$JAVAOPTS -Xms1600m -Xmx1600m"
1850 data points

The dev environment, running v3.5.6, has a very different free memory usage. Other than log entries occurring at the time of backup/purge, the mqtt connection has not been lost.

(You can see the influence of the backup/purge also in this chart)

I am concerned with the mqtt connection being lost as it is possible that data will be missed. This occurred once during a backup when the connection was lost for 10mins.

Is this issue due to the limited resources (memory) on the VM? I am not convinced it is as the dev environment does not show a similar behaviour. Is this issue due to v3.6.5 (compared to v3.5.6)?

Is this behaviour expected during a backup/purge? Can it be mitigated?

Any ideas would be great.

Thanks
Ian

PS: as another comparison, I have a similar app running on a physical box with 32gb memory, JVM size of 2.8GB and 2300 data points and have not seen this behaviour (mqtt connection lost) v3.5.6.

Hi @dgm

You are close:

            options="{
                synchronizeGrid: false, 
                valueAxes:[
                    {title: 'Connected', titleColor: '#ff4d4d', minimum: 0, maximum: 1, integersOnly: true, baseValue: 2},
                    {title: '°C', titleColor: '#ff4d4d', minimum: -50, maximum: 50}
                ]
            }"

Each axis is an element in the valueAxes array.

Ian

Hi @terrypacker

Thanks for pointing me towards the audit trial. I hadn't realized what elements were tracked here.

However, the audit trail doesn't offer any comparison tools, version tagging, etc that source control app would, like Github. This was my intention for saving the global scripts to an external file.

Thanks
Ian

Hi Matt, Craig

The functions I have created in the global scripts are being used by meta points and scripting data sources. I was not intending for them to be used by the mangoUI.

My hope was that I could move the global scripts and have them reside in a file(s) on the server outside the database. Currently, if I want to version control the global scripts, I need to export them using the configuration tool, save that file and then commit to source control. Comparing versions is a challenge as the scripts are each saved as a string in a JSON object. Any comparison tools shows the entire string has changed.

Updating (or creating) an app would then just require that I copy the file over instead of importing using the configuration tool. One less step to manage.

I'm hoping I'm making sense :)

Thanks
Ian