gRPC message exceeds maximum size

With gRPC active, you see error messages like the following in logs:

RESOURCE_EXHAUSTED: gRPC message exceeds maximum size

Cause:

Your content objects are larger than the default maximum gRPC message size of 4 MB.

Solution:

Add the following properties for client applications. These double the maximum message size to 8 MB. If you have even larger content objects, you might want to increase this value even more. The concrete message size in the error message will give you a hint on how much to increase the maximum message size. Keep the value in sync with the server configuration (see below).

spring.grpc.client.channels.cap.max-inbound-message-size="8388608B"
spring.grpc.client.channels.wf.max-inbound-message-size="8388608B"

Correspondingly, add the following property for servers. Keep the value in sync with the client configuration (see above).

spring.grpc.server.max-inbound-message-size="8388608B"

Applications do not respond on request

Applications do not respond on request. The CPU load of the applications is high, the thread dump shows threads that use nio classes.

Possible cause:

Problems with the CORBA ORB.

Possible Solution:

Add the following ORB property as a system property for the affected applications:

-Dcom.sun.corba.ee.transport.ORBUseNIOSelectToWait=false

CORBA Timeout for large copy operation e.g. when deriving a site

A large recursive copy operation like deriving a site fails on the client side with a warning like this after 12 hours:

org.omg.CORBA.COMM_FAILURE: WARNING: 00410019: Communications timeout waiting for response.

and the stacktrace (on DEBUG level) contains:

com.coremedia.cotopaxi.corba.CorbaContentConnector.copyRecursivelyToInternal

Cause:

The CORBA client ran into a timeout when waiting for the copy operation to finish.

Solution:

Increase the specific timeout for the copy operation (default: 43,200,000 (12 hours)) by setting the environment variable COM_COREMEDIA_CORBA_COPY_RECURSIVELY_TIMEOUT_MILLIS.

CORBA Timeout for other operations

A CORBA operation fails on the client side with a warning like this:

org.omg.CORBA.COMM_FAILURE: WARNING: 00410019: Communications timeout waiting for response.

Cause:

The CORBA client ran into a timeout when waiting for the operation to finish.

Solution:

Increase the global CORBA timeout (default: 1,800,000 (30 min)) by setting the system property com.sun.corba.ee.transport.ORBWaitForResponseTimeout.