Workaround for Bug # 62085 on ODS 7.14

Hello All Few weeks ago we started getting a condition where all application = sessions would hang (including dbaccess/isql) and the following string = would be written to the message log. listener-thread: err =3D -27001: Read error occured during connection = attempt. On calling Tech Support they told us that it was bug # 62085 and was = fixed on ODS 7.23.UC1. It seemed that the listener thread was waiting on = dead processes, and the workaround would be to bring the engine down and = up. Since our client was unwilling to upgrade, upgrading was not an = option for us right now. The first workaround we did was to try to increase the number of poll = connexions in the NETTYPE parameter to 250 on the shared memory = connexion (we have NUMCPUVPS =3D 1 so the number of poll threads is 1). = However all it did was to give us time before the error came up again = and we would have to cycle the engine at less frequent intervals than = before. Then one of our Unix programmers hit upon a better workaround. He = presumed that the shared memory implementation uses ipc calls and the = socket implementation uses rpc calls, so he assumed that the bug in ipc = will not be replicated in the rpc module. So changing the INFORMIXSERVER = to point to the soctcp NETTYPE may not result in the bug. So thats what = we did and we have not had the problem since! This is just for the information of anybody who may be having similar = problems. HTH Sujit Pal