Please Note This forum exists for community support for the Mango product family and the Radix IoT Platform. Although Radix IoT employees participate in this forum from time to time, there is no guarantee of a response to anything posted here, nor can Radix IoT, LLC guarantee the accuracy of any information expressed or conveyed. Specific project questions from customers with active support contracts are asked to send requests to support@radixiot.com.

Hi Stefan,

I assume you're talking about DGLux 2.5 (the one based on flash). Unfortunately, the answer is no, it does not backup for you. I've faced that problem a few times, and got more than a little frustrated. My various solutions are:

Is the DGML existent? If so, it may be salvageable! You can find the DGML in the Export item in the menu within DGLux. If you get a really large block of markup on your dashboard that won't render, there is one or more issues somewhere in the DGML. I usually found it via binary search: find the main container and elements and cut out half, import, see if it worked, If not try the other portion, if so replace it and try to halve that section. It's important to make sure you only cut out complete tags and their contents. It's kind of a pain. It'll end up being a malformed property of a tag, like some weird connector string but no actual value (sorry I can't be more specific, I haven't done this in a long while). Or it will all be empty, and you're out of luck. Go ahead and check it on the disk too, if it is empty in DGLux. It may still have all the markup in your Mango/web/modules/dglux/web/dgml (or other directories in dglux/web).

Which brings us to...

Vigilance. Save to your computer (via export in menu, I believe)! Save to server! Save everywhere. Shift + s

I got so aggravated when I was doing lots of DGLux work with this issue that I actually did all my development of new dashboards in a sandbox designed specifically to avoid this stuff. I had a timer element that counted to 300, and triggered a script, which increased a counter X and saved the dashboard (saving the dashboard is a behavior you can trigger) under a name like "autosaveX-dashboardName.dgml" It was irritating to have DGLux freeze for 4 seconds every five minutes, but it saved me three or four headaches over many months, so I accepted it.

It's definitely part of the reason we moved toward our angular dashboards and embracing standard web technologies.

Hi Tobias!

It is certainly possible! Unfortunately, it's not as easy as a click and drag. There is not a direct way to query our statistics engine from DGLux for binary points.

To do this, you'll want to use a Meta point in Mango to generate the information you're interested in. The meta point would look something like this, with p being the varName of the Door point in the script's context, running every minute (you can use the cron selector, or enter "0 * * * * ?":

statistics = p.past(MINUTE); if(statistics.data[0].value == true) return statistics.data[0].proportion; if(statistics.data.length < 2) return 0; return statistics.data[1].proportion;

Now we've got a data point giving us minute resolution (finer resolutions can be achieved in modifying the cron and call to past()) for the binary point. This data point will have the meaning "Proportion of minute open" and will be numeric.

From DGLux, you can query this point to get the values into a table. Then, you can bind the value column to a binding column on another sheet. You'll want to use the Sum rollup. Create another binding column on that table bind the value column again, and use the Count rollup. Create a formula column and have its formula be "return countColumnName - sumColumnName". Now you can bind these columns (sum and the forumla columns, click and drag the column headers) to a pie chart to get minutes in each state. If you wanted it to be a proportion, you could make formula columns and divide by the count. If you do not see anything in the sheet with the binding columns after binding (first step on the second sheet), click the add row button.

You can use the meta point's history generation function to get this to work even for your past point values.

Sorry I meant to say I change the data source to MODBUS IP because I could not get the COM to to work

^ You cannot configure it such that anyone may register a user. Only admins may create users, and this includes through the API. So, while you could have a portal to create a user built on the API, that portal would have to have admin permissions.

@phildunlap Very useful information. I searched and found exactly what I needed to know in this answer. Thanks.

It sounds like the websockets are not working which could likely be caused if you are using a proxy that's not configured for it. Can you share how you have the URL being directed to Mango. We use Apache with a virtual site to act as a proxy and had issues getting the websockets to work.

Here are some general instructions we put together that might be helpful:

Login to VPN server and enable Websocket proxy module in Apache (just has to be done once, don’t need to do it every time):
a2enmod proxy_wstunnel

Go to Apache websites configuration directory:
cd /etc/apache2/sites-available

Backup the website config file you want to set for Websocket proxy:
cp -pr <yourwebsite>.conf{,.bak}
Where <yourwebsite> - is the name of config file, usually the same with website name.

Edit the file <yourwebsite>.conf in your favourite editor and add below text:
RewriteEngine On
RewriteCond %{HTTP:Upgrade} =websocket
RewriteRule /(.) ws://<Mango_IP>/$1 [P,L]
RewriteCond %{HTTP:Upgrade} !=websocket
RewriteRule /(.) http://<Mango_IP>/$1 [P,L]
Where <Mango_IP> - is the IP address of a Mango device to which you configured ProxyPass.

Reload Apache configuration:
service apache2 reload

Hi Paolo,

It looks to me like you should uncheck 'configuration is in hex'. If you are translating a 30 into a 0 and 33 into a 3, you are not transporting in hex, you are transporting in UTF-8 or ASCII or some such. And '0x0a' is \n

You may be able to get this to work via regex like...

DATA SOURCE: Use terminator: Yes Terminator: \n Message Regex: (\d+;)*([0-9]{4})[0-9]{4}.* Point Identifier Group: 1 Configuration in Hex: No DATA POINT: Data Type: Alphanumeric Point Identifier: 0103 Value Group: 1 Value Regex .*0103([0-9]{4});.*

Also, If you are trying to access historical data via MySQL you would need to also uninstall the Mango NoSQL modules. On the MangoES and Enterprise Mango the historical data is stored in the NoSQL database which is separate. Here is a follow up video that explains more: https://help.infiniteautomation.com/support/solutions/articles/14000022517-enterprise-nosql-database

Note that by removing the NoSQL module you will not have access to your existing history. Also the MangoES will lose performance because MySQL is not as efficient at storing and retrieving time series data.

That was the issue and I have found the right path now.

When each will need to complete a poll durring intilization so it sounds like it's taking a long time for your slave devices to respond and for the poll to complete.

If you edit the data source you can see the last 10 poll time and will look like this:

0_1464974869733_upload-e1efb9d4-4536-42b5-befa-c5129c255eaa

This would be useful information to know. As Phillip mentioned using the new options for a multi threaded start up should speed it up but ultimately the issue is your slave device response time.

Once you have a license installed for Enterprise Demo which includes the NoSQL module it will no longer be a Demo and will be a fully working system.

Hi Mebius,

I looked at your logs, and a few things were interesting. Some of your polls, especially for your SQL data source, are extremely long. Even some of the PLC polls were up to twenty minutes it looked like, which is way too long. Do you have long timeouts with many retries? The SQL source has a poll that looks like it lasts 4 hours, which is also much too long. What kind of query is it running?

Do you have data in Mango during the times in which DGLux seems unresponsive?

Nice! Thanks for sharing what resolved it for you!

Glad to hear it!

Hi Tracey,

If your device is accessible via mangoautomation.net, it may have a vpn it belongs to. I don't believe we issue keys typically to clients to connect their own machines to the vpn hosting their device(s), but such a credential could get you past the need to gain local access.

Q1: Correct, they do not backup your DGLux files. You can find these in Mango/web/modules/dglux/web . You're probably most interested in your DGML files, but you will need your user and navigation files to restore it completely.

Q2: Correct. Stored in Mango/backup

Q3: iDrive is an option, and its one we like. Lots of the configuration and scripts are already in place on an ES to do an iDrive backup.

Q4: It is certainly possible. You would want to get some kind of controller for USB operations. You can check out something like usbmount and use it to trigger the copy of files you desire or if you're feeling more low level you can write a udev rule to kick off a script to mount/copy/unmount

Q5: "Should" is such a relative term. If you were collecting data on some classified widget, then surely the business protocols you operate under would govern if it were wise to let that data reside anywhere but behind your locks and your keys. However, since I'm assuming this is not a classified rock crushing application or anything so sensitive, then yes it is definitely a good idea. What's the probability of a rock crushing computers onsite vs. offsite? In general, yes, it is better to do offsite backups to wherever. The probability of a massive electrical event or general cataclysm affecting an area is greater than that event happening to two or more areas, even if still fairly small.

Q6: There are three backup options offered (counting NoSQL). 1. H2 backup - this backs up your entire database, 2. Configuration backup - this exports the configuration of the Mango to a JSON document, but has no runtime information at all (events, values). 3. NoSQL - this backs up only your NoSQL database. You need to have a database that matches the NoSQL backup to put Humpty-Dumpty back together again. Just restoring a Configuration backup to a new Mango will not have the actual point IDs guaranteed to be the same as the source system, and since the actual ID, not the XID, are used in the NoSQL module it is important these are retained. If you are running on MySQL, this would mean dumping your dataPoints table at the very minimum. Note also that if you were not using NoSQL the H2 backup would also backup your point values.

Q7: See answer to Q6, but yes, the H2 backup will also retain the data point IDs and therefore as explained in Q6 is necessary (if using H2) to restore both the point values and the configuration.

Good questions!

You are that man! It asked for my password after I enetered "sudo ./ma.sh start"...

Another fine JOB JOEL...

thanks for that I'll give it a go.

The point links solution still sounds good to me. You'll need an SSH data point for each RFID point I think. Then you can just hardcode the door number either into the pointlink or the ssh point. I would opt for the point like, and set the SSH point's value to something like "RFID# Door#" and stick with the ssh points' set commands being "key=$(/path/to/validation/script.sh VALUE)" as the string VALUE will be replace by our two arguments for the script.

Each RFID point will have a point link to an SSH point (and all SSH points can be the same, but for the sake of event handlers (if Mango needs to track state beyond just issuing a single open to the doors). Door 1's pointlink may look like,

return source.value + " D001"

You may need to check for an empty source.value and such. You can use all the same control logic from my earlier posts, involving another chain of point links from the SSH points to their relays.

Well, Joe, I'm making a dashboard in DGLux, and one of the things I want to display there is the number of nonsilenced nonacknowledged alarms from mango, so the user could see that on the dashboard, and go to mango to acknowledge or silence them.

Thx for the video and link.

@terrypacker Thanks for the update. Please keep in mind that this problem is not unique to my installation.