This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
>From: Michael Keables <address@hidden> >Organization: University of Denver >Keywords: 199911101751.KAA02013 LDM broken pipe Michael- Tom is out of town until next Thursday so I'll try to help you out. >Been having increasing frequencies of data loss on cyclone due to a broken >pipe in the XCD decoder: > >Nov 10 17:16:55 cyclone pqact[12767]: pbuf_flush (4) write: Broken pipe >Nov 10 17:16:55 cyclone pqact[12767]: pipe_dbufput: xcd_runHRS write error >Nov 10 17:16:55 cyclone pqact[12767]: pipe_prodput: trying again >Nov 10 17:16:55 cyclone pqact[12767]: pbuf_flush (4) write: Broken pipe >Nov 10 17:16:55 cyclone pqact[12767]: pipe_dbufput: xcd_runHRS write error >Nov 10 17:16:55 cyclone pqact[12767]: child 21245 exited with status 127 > >What does this mean, and how do I fix it? There are several reasons you could get this error: 1) The user LDM cannot find the executables. 2) The user LDM does not have permission to write to the file. Since things work some of the time (and assuming you haven't changed anything), there is another time I've seen this problem. I've seen it occur when the ldmfail script runs and does not switch to a different host very nicely. Are you running the ldmfail script from cron? If so, and if your connection to your upstream site is pretty reliable, I'd suggest commenting out the crontab entry, stopping and restarting your LDM. If you aren't running ldmfail, then I'm puzzled. Can you give us a login as the user ldm so I can take a look? Don Murray