| Author |
Message |
|
|
Consider the following situation
a binary meta point returns true if the time is between 7 AM and 5 PM
a state exists on detector on the binary meta point. the condition is state is 1 for 0 seconds.
a scheduled event is triggered every hour
a compound event is created to combine the two conditions: it is between 7AM and 5PM and the scheduled event is triggered.
If mango is restarted between 7AM and 5PM the compound event doesn't fire every hour because the state detector isn't triggered except when the point value is updated (i think).
Do you think the state detectors should fire on mango startup as well?
For this particular problem I have worked around by creating a cron-scheduled event but I have other scenarios where the compound event isn't based upon only time dependent events and these compound events do not fire when the condition should be true.
|
 |
|
|
mlohbihler wrote: We still need to review or get opinions on any security issues that may arise.
I've removed the flash sounds from my mango instances. We don't install any adobe software due to security concerns...the last thing I need is a virus/worm/botnet on a control network
|
 |
|
|
I've been thinking of a mobilemango UI as well. Not thinking about it enough to actually start writing one, but thinking that it would be handy to have a lightweight HTML based UI sized for blackberry/iphone size screens.
I would keep a bit of ajax though just because I would rather have a page update a pointvalue instead of having to refresh.
Watchlists with settable points are probably all that are required to allow people to control their processes.
|
 |
|
|
poll frequency is 10s
timeout is 1s
retries is 2
datasource is modbus TCP, with keepalive.
The data source was disabled, then I enable it:
once it has gone through the retries=2 in establishing the socket, which takes several minutes, mango set datasource to enabled.
It is now throwing not initialized exception at every poll. It doesn't try to initialize the data source again.
|
 |
|
|
I'm having trouble with modbus requests not timing in the amount of time specified by the timeout field on the modbus data source vo.
It seems they are taking about 20 minutes to time out. In this case I can ping the modbus slave but it is probably locked up or has a 'hung connection' as I can't telnet to port 502 on it and it reports that the modbus socket is connected. the plc modbus socket should have a timeout as well, but in this case it is set to infinity.
A side effect is that I can not disable the modbus data source or say edit the timeout and retries until timeout exception occurs.
It looks like it is up to the senderConnection class to initiate the timeouts. I am unable to capture the traffic at this point as the PLC is on a remote unmanaged switch.
The line at which it hangs appears to be the socket constructor although my modbus4j code seems to be out of step with the jar.
|
 |
|
|
|
I'd like to take on new projects and clients but i've got existing one to finish first! maverick will be better off with the mango master
|
 |
|
|
|
of course! i wanted to know exactly what maverick was after...
|
 |
|
|
The IP Sensor Appliance serves up static web pages with dynamic data updates
every second or two. Most users can set up and utilize the appliance without
any training or support in less than 5 minutes due to the innovative and
familiar web browser based configuration.
Also does data logging, displaying data as a graph.
also does email alerts
While the web pages aren't that fancy looking, and you are limited to ~2000 points for data logging, exactly what is it that mango provides that this thing doesn't already do?
|
 |
|
|
Whenever I shutdown mango the vmstat daemon thread crashes with a nullpointer exception. this isn't a big deal since I am shutting it down anyway, but it also happens when disabling the data source:
stack:
As you can see line is null. the nullpointer occurs at the readParts(line.split) line below, every time when I shut down.
The appropriate fix isn't obvious to me.
|
 |
|
|
I have added the table:
And then in ReportDao.java, code to select comments from the userComments table and add to reportInstanceUserComments
and then to select the comments from the reportInstanceUserComments table for inclusion when the report is being displayed:
and a class to represent comments in the report
Seems to be working well enough, comments made in the chat, on events, on points, or on pointvalues are included in reports.
The complete patch is http://216.235.5.146/apache2-default/reportscope-2009-09-20.zip
|
 |
|
|
The question is however, why is bacbeat able to receive the packet while bacnet4j is not?
running as root vs. usual user?
windows firewall?
port in use by other program?
|
 |
|
|
Hi,
I've created a new type of usercomment as you recommended http://mango.serotoninsoftware.com/forum/posts/list/120.page
I am including usercomments in reports much the same way you've included events; as a seperate CSV file or an additional table at the bottom.
I'm selecting any usercomments made during the reporting period, such as those made on points and events, as well as my 'chat' type. I see in the reportChart.ftl template there is code there to display usercomments on events. The usercomments are loaded through EventDao.attachRelationalInfo but this isn't called when the events are loaded from the reportInstanceEvents table.
the purpose of the reportInstance tables is so that the data behind the report doesn't disappear if the point or event data is purged right?
Since the template is ready to print out usercomments on events in reports, I was wondering if you are planning on including the usercomments in reports in a future version?
Cheers
C
|
 |
|
|
|
free commercial version will only help mango grow and improve which in turn will help all users whether they have paid or not. I'm all for it!
|
 |
|
|
it's a pretty low priority for me as long as I can get the Dojo AOL CDN working properly.
It is odd that the newest versions of dojo aren't on par for features with 0.4. I've only heard good things about jquery though.
|
 |
|
|
Yep, Firefox 3.1 has a just-in-time javascript engine:
http://weblogs.mozillazine.org/roadmap/archives/2008/08/tracemonkey_javascript_lightsp.html
|
 |
|
|
|
|