Usage was staying between 80-99%, which began two days ago. Nothing has been changed (that I’m aware of since December).
So far, in the forum I found where it may be helpful to add a line to the Garbage Collector adding a slight pause and reducing the frequency of operation. (I will do this tomorrow as its in a production environment)
I have tried to look at the logs, but not sure exactly how to run down each item as the web portal shows gateway-shared-exec-engine-8, id=55 (not just this one, but there are several different listings that show “blocked”
Also, MVStore background writer was using almost 20%. Then disappeared. I will upload the wrapper log, thread dump, and config.
Several devices have been disabled as they were faulted and usage has dropped some, but I would like to see it much lower. Any help would be greatly appreciated.
dump.zip (17.2 KB)
wrapper.log (2.3 MB)
If you’re already planning on downtime, I would strongly recommend upgrading to the latest 7.9. 7.9.5 is years old, and has many bugs of varying severity.
I don’t get the final say, unfortunately. I was called over a few issues back in October with the portal showing “internal error” after clicking on anything on the page. It would work for 1-2 clicks, then Internal Error. I recommended that they upgrade then based on advise here, but they obviously did not listen to the warning/advice.
I will lobby for the upgrade, but expecting the worst I’ll keep looking for an answer.
-Forgot to include in my original post that most of the usage is in the “Unclassified” section. Devices/tags/web are using less than 15-20%.
Your thread dump shows an astonishing number of socket connections and netty event loops for drivers burning CPU (runnable state). I recall seeing this once. Pretty sure it was a netty bug. You need to upgrade.
Note that there’s never any “hot-patch” or similar bug handling for Ignition. Fixes for confirmed bugs will always require upgrading.