Hi Mariusz & others,
I would recommend you try upgrading the OCP Cluster to a OCP v4.6.12+
version.
We saw a significant decrease in Idle CPU Consumption with the EFK Stack
(cluster-logging) with OCP v4.6.12. The reason is most likely a patch for
the increased default logging level (Bug 1895385 -
That is progress!
About network:
Are you using vswitch in z/VM?
is isolation turned off?
Is there any monitor to show network activity, like z/VM Performance Toolkit?
At least you can use TCP/IP netstat command to see high level activity.
How many OSA and chpid are in use? What is their speed?
-
Hello,
Unfortunately I won't be able to send monitor data due company rules. I
started with reducing the number of vCPU and this indeed helped with
application performance. I also increased SHARE REL for worker nodes, where
application is running and set CPUPOOL CAP limit on infrastructure node
(w
We're making progress with dotnet on Z. Lots of work still to be done but Uli
Weigand has been making headway in leaps and bounds and the dotnet keepers have
been superb in their support…
dotnetbld@aussie-3:~/HelloWorld$ ~/dotnet/dotnet new console
The template "Console Application" was created