Here is a pic of a freeze
here is a pic of memory in the console within 30 minutes.
is it possible my memory is set to low and im maxing out causing freezes?
anyone seen igniton screens do this?
its happened to 2 different HMIs, happens at random usually early morning, and within the last hour.
Thank you
Almost every time I’ve investigated such problems, they turn out to be programmer error. Specifically, doing a something in a foreground script that makes a gateway round trip. Anything that perturbs that round trip changes an unnoticeable milliseconds-level pause into a very noticeable multiple-seconds pause.
Audit your event scripts and runScript bindings for system.tag.*, system.db.*, and system.alarm.* calls. The only exceptions should be button actionPerformed events, since the user expects the system to go do something at that point.
This only happens on two HMIs running PIs mirroing off a VM.
and the only scripting added recently by my self you have seen on my other post. And I don’t think they relate to those 3 things. Maybe they increase the prior issue? I really don’t know at this point.
IT says they see the session disconnecting. But we are not sure why. And if it was the recently programmed code, why not on the other 6 clients?
Try stopping the client and deleting the client’s cache. I have seen this happen too. Sometimes the files in the cache get corrupted and cause the client to get very slow and use a lot of memory.
No, when the session, that for vision is actually a client, gets disconnected, a notorious message appears in the disconnected context.
Do a control+F and search within all windows/templates (make sure everything is selected to be sure) for runScript and then within those, look to see if the functions that are being called call system.tag\db\alarm.
If you find nothing, start searching for system.tag\db\alarm one by one - it’s possible they exist in some property change script that is perhaps taking much longer than you think, or worse getting stuck in some property change infinite loop.
If a script uses multiple runNamedQuery calls, would the round trips be mitigated by calling the same script through a gateway message handler?
No. Maybe. If using .sendRequest(), no. If using .sendMessage() to gateway scope, and listening with a client message handler for the response, yes.
When I search via ctrl + F the the desinger freezes, doesn’t search and i have to end it via the task manager.
I also see a faild backup around midnight then everything restarting.
12:52:51 AM ScheduledBackupManager Scheduled backups config error: unable to create directory "\\abcdefg\files\BKUPs\Ignition".
I don’t think this would cause a client to freeze.
Regarding Control+F, try opening the problematic window(s) one at a time, and looking manually for runScripts, or property change events that call system.tag\db\alarm.
If you’re unable to even look through the designer to find what is wrong though due to it freezing up, then it’s probably time to call Ignition.
@pturmel I have been looking and cant find any scripts like that. everything is actionperformed or indirect tag etc…
any other ideas?
I think you should call support.