Before you use real-time runtime metrics collection, ensure that your Linux® environment can handle the data flow from real-time runtime metrics collection.
About this task
You can take the following actions to ensure that network
performance on your Linux environment performs
optimally.
- Avoid Domain Name System (DNS) delays. This problem
can manifest on z/TPF as a timeout condition on
your first message analysis tool collection and potentially
succeed on following collections.
- Avoid overruns on the Linux
socket receive buffer. This problem can occur on
the z/TPF system in the form of a timeout condition
on real-time runtime metrics collection data types.
Procedure
- To avoid DNS delays, verify that your DNS settings list your preferred local DNS servers before public DNS servers.
- View the contents of the file:
/etc/resolv.conf
- If your preferred local DNS servers are not listed first, modify the contents of
the file to list your preferred local DNS
servers first and save the file.
- To prevent Linux
socket receive buffer overruns, verify that the
current receive buffer allocation of your socket
is at least 4 MB or greater.
- If you use Docker, take any of the following actions:
- If you use another system type that is not Docker, complete the following steps:
- Navigate to the directory where the system configuration files are by entering the following
command. Ensure that you locate the correct system configuration file that your environment uses.
cd /etc/sysctl
- Edit the system configuration file to update the value of the socket buffer to 4 MB or
larger.
- Verify that the socket buffer value is updated.
- If the socket buffer value is not updated, contact your IT manager regarding allocating
resources.