McGuire_Ed@GRINNELL.MAILNET.UUCP (04/11/87)
I'm describing a problem we're having that looks to be a long time in the solving. Have any of you seen (and perhaps solved!) this one? Users are complaining that programs occasionally abort with an error traceback that says a timeout error on SYS$OUTPUT was the source of the problem. These are images such as TYPE, or our mailer or news system, that send large bodies of output to the terminal. If the output is larger than the terminal's internal buffer, and the user presses the Hold Screen key to pause output, and waits a couple of minutes to continue, the error occurs, roughly 50% of the time. We sat on terminals and reproduced this error rate consistently. The error occurs on all DMZ-32s attached to an 8600, with a MICOM Micro-600 switch between them and their DEC terminals or microcomputers (doesn't seem to make a difference what kind of terminal). It doesn't occur when the interface is an Emulex DMF-32 emulator on an 11/750. DEC Field Service said something about VMS development not getting their DMZ driver ready in time to distribute so they hacked an interim driver out of the DMF driver. He claimed the buggy driver was responsible. When I called Colorado to see what they thought, they said every time people call to report timeout errors, they forward the problems to the local office because they're so difficult to diagnose. I'm waiting for the local office to do something. Thanks in advance for any suggestions! Ed McGuire Grinnell College MCGUIRE@GRIN2.BITNET