6 Replies Latest reply: May 15, 2014 10:13 AM by Olly Bond RSS

    Monitored Process Setup Error

      Here’s the challenge:

       

      I have one Data Pump process with two projects.  The first project that runs has a file as input.  The second project has a database query as input.

       

      When I manually run the job, it’s great.  However, now that it’s completed, I would like to schedule this “multiple projects per job” to run when the file exists for the first project.  So I defined an input distribution for the first project, included the input file on the monitoring tab and set enable. 

       

      However, when click OK and attempt save I get the error message: 
      path\2nd project.xprj does not have an input distribution defined.  This is required for a monitored process.

       

      I have the first project set to have an input distribution, however, the second one is a database query, so it does not have an input “file” and will not allow an input distribution to be set up.  We tried to trick the system by clicking text, and setting the input distribution, then clicking back to the data base query, but to no avail.

       

      This, to me, seems like something you would want to do on a regular basis, that should be standard.  I want a monitored process and as soon as the input file for the first process is finished and the input file “distributed” I want the next job in the process to kick off just as it would normally do.

       

      Any ideas?

      Thanks.

        • Monitored Process Setup Error
          joey

          I've seen a few posts here about using scripting to manually kick off a process at the end of one process. If you can't find anything, I can look. you'd need to split this into two processes. The first project would be monitored. The second would be an ad-hoc (no schedule), and would be kicked off through code at the end of the first process.

            • Monitored Process Setup Error

              Thanks for the response.  I was afraid that I would have to go with the scripting option.  I was hoping that someone would have a “better” solution.  I worked with the forum a year or so ago on a scripting issue involving reading the log so I’ll go back and start there along with a thorough forum search on scripting and executing the pump via script.  If I end up with a workable piece of code I’ll post it to this thread.

               

              I have to say that I think this is a major flaw in the way monitoring is designed in that monitoring is enabled at the process level yet input distribution is satisfied at the project level.  It seems to me that monitoring should be totally separated from input and only exist at the process level.  Unfortunately Monarch support is not an option here as every time in the past when the solution has come down to scripting they refuse to provide that support.

            • Monitored Process Setup Error
              Chickenman _

              John,

               

              You say, "the second one is a database query, so it does not have an input “file” and will not allow an input distribution to be set up."

               

              It seems (and I just tested that this can be done but YMMV on whether it will "work") you can invoke the XprjEditor from the second project and choose the Input tab.

              Then  and navigate to the location of the database (use File Type: *All Files) and choose the .mdb, .xls or whatever it is.

              This can then be given an Input Distribution of "Delete"

               

              CM

                • Monitored Process Setup Error

                  Chickenman,

                  Thanks for the response.  In my case the Input tab “Select a data source” is the database connection, the “Table:” is a database table and then I’ve added a filter.  That’s what I meant by the project not having a file as input – it has a database table as input.  When in this mode the “Input Distribution…” button is greyed-out.  If I could get to the input distribution I could add a log message distribution and it would satisfy MDP’s need for an input distribution (this is how I got around the problem with another process with many projects using the same input file – I do a log distribution on all the projects except the last one).  Given that, I don’t see an .  Am I missing something?  Thanks again.

                    • Monitored Process Setup Error
                      Chickenman _

                      I mis-interpreted how your Projects are working. My assumption was that the first Project ran and the output is an export like .mdb to some location, so then the path to that table could be Added as the Input for the second Project and in that case the Input Distribution button is active. Obviously that's a bad assumption.

                      But if this is a static DB table sitting soemwhere on the server or network then monitoring wouldn't work because it would always be present and therefore "on" and Log wouldn't help anyway.

                       

                      Forgive the confusion.

                      CM

                        • Monitored Process Setup Error
                          Olly Bond

                          Hello John, hello everyone,

                           

                          I think there's a way around this - to run one process as monitored, and to have a post process script that then calls a second ad hoc process which consists of the next project.

                           

                          I'm not in front of Datapump to check or to give you any code, but I hope this is helpful.

                           

                          Best wishes

                           

                          Olly