I have a proc-test that is failing for one node because proc-test data is truncated. The server is configured with the default limit of 512K for (MAXMSG_CLIENT). Looking at the client_data for the active test, this is showing a complete ps-list, but it also seems like the total of client_data is over the limit of 512K. Looking at a historical test, it seem like the client_data is truncated at 512K and it matches the same spot in the ps-list.
An obvious solution for this might be to increase the global parameter MAXMSG_CLIENT, but will be valid for all clients. Another possible solution, but not wanted, might be to change the client behavior to local handling.
A third solution could be to reduce the amount of data sent by the client. Ex stop the netstat output, which for this client is quite long and something we don't generate alarms for today. Does anyone know if this is possible, and if so how? Anyone that has an better suggestion on how to solve this?
Br, Øyvind