I haven’t personally encountered this issue before.
Are you historizing your gateway CPU usage tag? Located at [System]Gateway/Performance/CPU Usage. If so, does it’s trend match what you see on the gateway status?
Yes, it’s the same value there and It doesn’t change, but should. I still can’t find the issue
For an issue like this, I think it’s always best to reach out to support so they can look for themselves. Otherwise, I’m sure a gateway restart will do the trick.
Hello guys, I update the version to 8.1.20.
The frozen was solved and in another case the “-100” in cpu too
thanks !
Have you tried restarting gateway? (service)
So, yes I have and that fixes it but, 3 or 4 weeks from now it will happen again.
I have the same problem. After reset gateway, a month later the CPU trend frozen. Everything seems to run OK. The CPU cannot be constant as the garbage collector is still working up and down, should the CPU also? Looks like is ignition problem after accumulating 1 months of data.
We are on 8.1.37 and still have the exact same problem. Everything seems to be working normal but the CPU trend freezes about a month after restarting the gateway.
@here, we're tracking this internally as ticket IGN-8636, but be aware that this is an "upstream" bug in Java itself. So far, we've only noticed the issue occur on Windows hosts, and primarily in virtualized environments. Anyone running into this issue is encouraged to contact our Support team, ideally with the following information at hand:
- OS type/version?
- VM software and version?
- Microsoft Hyper-V Version (if applicable)?
We've got an active ticket with our Java vendor, but the more information we can provide them (and the more reports we have of this occurring), the better.
3 Likes
When that CPU usage gets 'locked' at a certain rate (like 0.8%), is it actually limiting its usage, or is it just not updating (but still using more in the background)?
We have a few ignition servers that keep hitting this issue recently. We have a perceived "slowdown" / timing delays when it occurs, but also this morning we had an issue where a specific application wasn't executing tag change scripts while that occurred (may or may not be related, but error logs didn't say anything).
There isn't actually any limiting of CPU usage in this scenario--rather, the inquiry that Ignition does to the JVM to show the usage % (which then calls into native code to retrieve the information from the OS) is not behaving properly.
1 Like