I’m currently working on a project based on a Desktop PC running the Ignition GW, and 9 Industrial PC as client.
For each of those 9 client, I have a TV Monitor displaying the Ignition Project page, showing a data table, sometime a picture and a bunch of tags.
The Clients have no input, no keyboard or mouse are present, they just show data.
The PC Server is a x64 with the java version “1.8.0_181”, the same that I installed in all the clients, except that the client mount x64 processor but a x32 Windows 10. ( Actually one of them is been updated to a x64 Windows 10 )
After few hours of works, those clients seems to freeze, no input or network interface response, just freeze with the current state of the project screen, no errors, nothing at all, just completely frozen, until i force a power off by keep the power button pushed.
Initially I thought was a PC Hardware fault, but now I noticed this happened randomly on all the 9 PC, so I’m pretty sure this can not be a hardware problem.
Then, as i mentioned, i uninstalled the Java 10 environment and downgraded all the machine to the version “1.8.0_181”.
This problem happened on the client who got the Windows x64 Version as well.
Could someone help me out understanding what’s possibly going on?
Generic UI freezing, in my experience, has always been traceable to some form of infinite loop. This can be non-obvious, like a propertyChange event setting another property, which makes a nested propertyChange event, which then sets the first property. Or propertyChange events not checking the propertyName, so they respond inappropriately.
Generally you have to add logging to all of your event scripts. Or event counting with timed reporting. Watching the diagnostics console when a client freezes might help.
Thanks for share your experience with me pturmel,
I will look for any possible I could find, my only concern is the page that cause the freeze was really simple, with a table polling data from a database, and two buttons with only the “visibility” property being driven from a tag.
Have you verified that the time it takes to run the query is less than the polling rate of your table data?
No I did not.
Could that be the cause for a total freeze?
That could make sense because thought, I will try to increase the polling time, fortunately in this case I don’t need a quick refresh ration on the monitors.
At the moment it is set on the default value, do you think if a congestion on the network happen, continue request could be drive to a freeze of the client?
I’m pretty sure that in normal conditions, the time necessary to retrieve the data I asked for, it should be more than enough, my concern is what could happen is cases when the network, for any reason, get slower.