I have one app, running on two clients, and they have both locked up during update pushes. One PC is an onlogic, with 4GB, the other a thinkcenter with 8GB. I did find that the client setting in the project was default and set a 256MB. I raised it to 1GB start and 2GB. Attached is a picture of the client performance. This app has about 4000 tags, very limited dB queries. The most intensive screen has 18 spark lines with 10 second updates (slow levels) and 2 easy chart realtimes with 1500ms updates of 4 pens each. My deadbands are all set on historicals properly, set to discrete not analog or auto, so no interpolations.
Im hoping that since the normal memory peaks are around 600MB, it was simply running out of memory with the 256 MB cap. But looking at the graphs is there anything else i should be looking at? Ive tried very hard to follow all the best practices. Device drivers are running 0% overload. I dont see anything in the logs as far as issues of bad tags, or crashing scripts. I dont run any gateway scripts, and minimal client scripting.
Vision 8.1.38
Did it again....client locked up, and pegging cpu at
100% at 2GB memory.
This time, i kept the client performance trend running. When update became available, client and performance screen kept updating, cpu was at 2%, etc. But the performance trend and client are completely unresponsive. I had to drag task manger over just to see it as i could not minimize the performance trend. Keyboard, mouse, etc all work, but cannot close performance trend, and nothing within the client is responsive.
I would reach out to support, but also, the server stats probably have nothing to do with the client stats. I've ran much larger projects than that with the default client memory settings. It looks like you have crowdstrike and you might see if there's any antivirus exclusions you need to add as it's possible it's interfering with the updates.
Thanks for the reply. Im starting to think it does have something to do with ITs lockdown of this PC. Unfortunately, outside my control, as im just an integrator and this plant is locked down like fort knox. But considering this PC spent time on the business network before it was repurposed for ignition, im guessing theres some residual policies effecting Ignitions attempted updates.
In the client log viewer are there any messages that you can find in there, or is the diagnostics page basically locked up once you push the update?
Locked up. I can use windows keys like alt-tab, ctrl-esc, but the client and in this case the client diagnostics retain the upmost z order, so anything like task manager appear behind. In other instances the CPU and the zulu process peg at 100% and full memory usage, to the point of where its difficult to even open task manger. Killing the zulu process, restores the PC and relaunching the client, it opens with the updates. Today, it ran all day, but i did not push any updates. Memory on both clients cycled around 600MB.
You could look and see if there's any relevant log files in the user profile directory %userprofile%\.ignition or any subfolders, but otherwise I'm sure support will be able to help the most. (Or a fresh reload of Windows LOL with an activation this time (noticed the Activate Windows overlay)).
Lol...yeah i dinged them on that too.
