"Failed to read reply" for Tech Services -- with multiple users per port
Description
Cataloging and Acquisitions GUI users receive the "Failed to read reply" message when there is more than one person using the pc_server at a time. The problem occurs most commonly in connection with timeout, but occurs at other times as well. If there's just one person using pc_server_6nnn, the error doesn't occur.
There are no messages in the pc_server log at the time the error appears on the PC; it's a communication problem.
None of the causes described in the article Failed to read reply" in GUI **MASTER RECORD**, including end-point protection software or campus IT limitations on certain ports, applies.
Resolution
Extensive analysis using wireshark on both the PC and on the server failed to determine the cause. A successful workaround was to start a separate pc_server for each Technical Services user -- in this case, 20 separate pc_servers on 20 different ports.
Note: The "setenv PC_NUM_SERVERS" (the number of child processes for each pc_server) in $alephe_root/pc_server_defaults was reduced from 5 to 3; the idea being that with just a single user per pc_server, there would be less need for child processes. This worked OK.