Messages posted by: skiv71

would this be hard to implement?

be a great feature... watchlists, data sources, views....

thanks

skiv

this browser has always been my choice for using mango, but in recent years, revisions have made mango less functional....

i.e. alarm banner at top used to flash, doesn't anymore...

now i'm seeing errors in the logs where google chrome clients connect...

WARN 2012-04-25 12:56:49,828 (com.serotonin.mango.web.dwr.MiscDwr.jsError:193) - Javascript error
Description: Script error.
Page:
Line: 0
Browser name: Chrome
Browser version: 18
osName: Windows
location: http://xxx.xxx.xxx.xxx:8080/login.htm
WARN 2012-04-25 12:58:10,750 (com.serotonin.mango.web.dwr.MiscDwr.jsError:193) - Javascript error
Description: Script error.
Page:
Line: 0
Browser name: Chrome
Browser version: 18
osName: Windows
location: http://xxx.xxx.xxx.xxx:8080/views.shtm?viewId=6

does anyone have any thoughts?

no such errors with IE8 clients (but i hate this browser, too slow).

think i'm using 1.12.4

thanks

skiv

you should look into a suitable modbus gateway...

in my experience, i choose a beijers/mitsy e terminal for transparent modbus/plc connectivity...

anything it can talk to, you can

hi sorry, think its 1.12.4

hi, i use google chrome when constructing mango projects, and i've noticed what maybe a bug when creating mailing lists?

at the moment i'm having to revert to IE (which is dog slow) for this part...

When i create a new mailing list, i cannot overwrite the XID for it, i cannot select users from the drop down list, its just like the page has crashed.

In IE i can do all the above, but i still cannot drag over text to make it blue (and then replace-overwrite it).

Bug in the 'mailing_lists.shtm'??

Hey Matthew,

Long time... how's things?

That's a real shame, the beauty of the individual accounts and web access is the independant nature of things...

Just wasn't right, logging on whilst i knew there was an active alarm present but the event logger says no, because another operator ack it.

Not a problem in a single instance, viewing platform, even if it is VNC'd, but clearly is in a multi login web environment :-/

Hi, been using mango for some time now and spotted something i'd like to see handled different.

Not sure if this is just a setting, but in a multi user environment, it would be very useful if alarms were acknowledged on a per user basis.

For example, local instance on a site, operator acknowledges a fault, but the fault still exists..

I login after the operator has acknowledged this fault, and to me (and any other user) it will appear as if there are no active alarms.

Gives the wrong sense of a systems state.

I see on the options for users, there is a tick box for own audit events, how about ALL events?

thanks

skiv

hi, not sure if this is the right forum, but here goes...

i have some meta points which SUM a real data point (modbus) with a virtual data point (used to provide an offset for calibration purposes).

Because i needed precise update timing (for my data logging), i used to CRON to produce the result.

What i would rather have done is have the meta point update on context and use the polling interval of the real data source as the trigger (i have quantize) selected.

thing is the internal data point will also cause an update and this is undesireable.

it would be nice if i could enter 0 update period or disable it in some other means.

i thought of entering a very long update period, but thats no good.

really need to disable periodic update.

any thoughts?

thanks

skiv

i've decided to make 2 data sources...

a/. source with 1 sec polling for realtime data

b/. source with 15min polling and quantize checked

source b is set for logging on time stamp so achieves the desired result.

however, i'm also publishing this data (potentially both sources) with another scada instance retrieving the data.

currently the publising data source is set for updates by changes.

since the source is quantize, it would be nice if i could select 'timestamp' change for the update.

presently i only see option to select 'all changes' which see's the client scada have per second updates, or 'changes only' which fails since the data value might not have changed.

i'm using 1.12.0, has the publishing facility been changed to send updates based on timestamp?

thanks

skiv

hi,

i have a data source which i'm looking to set update on a per second interval.

but i would like to log data every 15 mins.

when producing reports, is it possible to get the 15 min interval quantize, i.e. 0min, 15 min, 30 min, 45 min etc etc.

i have managed to get 1 sec polling with 15 min interval logging ok, but i really need to have it on the times stated, not 15 min elapsed.

for now i've changed the data source polling to every 15 mins and logging to when time stamp changes, but this is not really practical.

live readings need to be update more frequently.

thanks

skiv

i cannot use event triggers based on time because i need them to be dynamic... but an alternative would to have access the elapsed time of a trigger to i could see where its got to... can i do that?

hi,

i'm using a binary flag to decide whether to increment or decrement a numeric data point.

i.e.

value=readlast.value;
if (bit.value==1) value=value+1;
if (bit.value==0) value=value-1;
if (value<0) value=0;
if (value>60) value=60;
return value;

i've tried this on context change, but because one of the context items is actually the output, i get false triggers (tried execution delay but dint help)

so i resorted to cron, 0/1 * * * ? which does work but i'm also getting odd behaviour.

but i've noticed this from the watchlist.

i'm polling every second from the main data source and watching the seconds counting away, it quite often jumps ahead (on totally unrelated sources and my point) then levels out.

so on the whole, 60 secs is 60 secs, but it will jump from 13 - 15, then hold off etc.

whats responsible for the internal timing mechanism?

seems bit erratic :-/

hi joohwan,

sorry i'll try and break it down

mango (since 1.8.1 i think)

modbus I/O, RTU via a TCP gateway.

all i know is it appears mango's polling mechanism is not geared towards multiple sources, i.e. there is no staggering of devices... and i've seen on site where this has risen to unrecoverable comms downtime. resetting the data source or restarting the instance is required.

2 other modbus masters I have at my disposal, seem to polling 'one at a time' and this works much better

hi,

i have a couple of RTU based I/O units which i reach via a MODBUS TCP/IP 485 gateway device.

using another modbus master and a polling interval set to say 1 sec, you can see the COMM LED indicators cycling nicely and all is well.

if i then disable that master and try mango as the MASTER, with a 1 sec update, the polling is very erratic and quite often misses the beat/rhythmn and shortly after comms falls over, with data source might not be updating etc.

i then have to disable and enable to source.

i never did get to the bottom of this but as it was the mango wasnt resilient enough to use.

but i have need to use mango this time so need to nail this issue.

any pointers as i try to revive this and repeat the problem?

could the polling be compromised by other sources being lazy?

if this was the case, i'd still like to see datasource polling by cron, that way the user can segregate sources completely.

thanks

skiv

hi,

i have a source which polls once every 60 secs...

i have alarm limits which trigger process commands (batch file) via event handler..

the external device which is communicated with via the batch file enters a busy state during communication.

my scripts etc cater for this by port checking..

problem is, due to polling, if 2 event handlers trigger at exactly the same time, neither one really gets a head start on the other, so neither sees the port busy.

basically i cannot handle simultaneous triggers...

is there a global setting for staggering process commands? 5 seconds should do it.

thanks

skiv