DSXchange

The AnalysisServer says the process is running when started and then when I do a status its not running and produces nothing in the MGXPgmInit.std.log or MGXPgmInit.log for me to look at......I beliebe my paths are right and the MessageSwitch is running just fine....any ideas? Almost like its runnin...

I have a Oracle Account called "psdb" which has the proper Oracle priviledges and works just fine logging into ProfileStage UID. But when we create other non "psdb" users they are failing with ORA-01031 or insufficient priviledges. These users have the proper role, priviledges an...

Yes PeopleSoft E1 is best but you can always ODBC....

Just do a netstat -a | grep ds and you should see the dsrpcd in a LISTEN state on the port.....or as others said if you can login then it is Listening....or if your DataStage Admin then do a uv -admin -info.....But otherwise nothing per se "hangs"

I gave my opinion....no more no less.....

Well a Hash is the file that is used for lookups in a Server Job....Parallel Jobs uses either FileSets or DataSets via a Lookup Stage

You can run dsjob from the Unix command or DOS command line.....You can create a Unix Shell Script or a DOS BAT file and run it there.....You can have a Scheduler (e.g., Autosys) call the Shell Script and so on....

Well before panic and corruption....try reindexing the Project.....

If you don't bring DataStage Engine down properly such as users still logged in or running/orphaned PIDs then you get this behavior. If DataStage comes up successfully and you get this error. Do a netstat -a | grep ds and the RPC daemon port should say LISTEN and if it doesn't then the port is grabb...

Seems like you were trying to upgrade with DataStage not being stopped properly....PID's orphaned is a indicator.....

Well a 0 means commit after all rows are processed or one giant commit....when too many rows you can see this is not a good idea....for small dimension tables its typically done....

Best practice is to set Array Size high to lets say 5,000 when processing large amounts....

Well first IBM/Ascential has not determined officially that "Hawk" will become version 8.0..... Second: 1) The Hawk version of DataStage will support EE (Parallel) and Server versions. 2) New customers will be provided EE and Server will not be presented stand-alone.....It's supported for ...

Have the DBA increase the amount of extents or better yet increase the "time" so the snapshot is not "too old"

The unlock suggestions are good but just to close the loop you may in addition have phantoms left out there.....even removing locks may still give you the error of exclusive so look for phatoms....

ps -ef | grep phan

kill -9 each phantom

If your running via dsjob then set the parameter of -warn to 0