I have been having trouble with my newly upgraded PeopleTools 8.51 environment. The problem was that I had not set the PS_CFG_HOME environment variable plus I also did not have the report node configured for the Process Scheduler server definition.
So, my process monitor was showing everything hung in “Posting” status:
I found the error messages in the c:\documents and settings\\psft\pt\8.51\appserv\prcs\HR90DMO\LOGS\stdout.
=================================Error=============================== Message: Database error encountered Info: Section: Info: SQL Stmt: SELECT R.PRCSINSTANCE ,R.ORIGPRCSINSTANCE ,R.RECURORIGPRCSINST,R.MAINJOBINSTANCE ,R.PRCSJOBSEQ ,R.PRCSJOBNAME ,R.PRCSNAME ,R.PRCSTYPE ,R.RECURNAME ,R.TIMEZONE ,R.PT_RETENTIONDAYS ,R.PTNONUNPRCSID FROM PSPRCSQUE R ,PS_PRCSRECUR S WITH (READUNCOMMITTED) WHERE ((R.RUNSTATUS IN ('9', '17') AND S.INITIATEWHEN = 1) OR (R.RUNSTATUS IN ('3','6', '7', '9','10','12', '17') AND S.INITIATEWHEN = 0)) AND R.INITIATEDNEXT = 0 AND R.RUNLOCATION = '2' AND R.RECURNAME > ' ' AND R.PRCSJOBSEQ = 0 AND R.RECURNAME = S.RECURNAME Explain: [Microsoft][SQL Server Native Client 10.0]Communication link failure ===================================================================== =================================Error=============================== Message: Database error encountered Info: Section: Info: SQL Stmt: SELECT Q.PRCSINSTANCE ,Q.JOBINSTANCE ,Q.MAINJOBINSTANCE ,Q.SESSIONIDNUM ,Q.OPRID ,Q.OUTDESTTYPE ,Q.GENPRCSTYPE,Q.PRCSTYPE,Q.TUXSVCID,P.PRCSOUTPUTDIR FROM PSPRCSQUE Q ,PSPRCSPARMS P WITH (READUNCOMMITTED) WHERE Q.RUNSTATUS = :1 AND Q.SERVERNAMERUN = :2 AND Q.PRCSINSTANCE = P.PRCSINSTANCE Explain: [Microsoft][SQL Server Native Client 10.0]Shared Memory Provider: No process is on the other end of the pipe. =====================================================================
I knew about the PS_CFG_HOME variable, but I had just skipped it in my rush to just get something working. What I didn’t realize is that it would prevent me from seeing even an SQR. So, I fixed my PS_CFG_HOME variable:
I also copied the files from the appserv directory under documents and settings to the one in my PS_Home where I configured my PS_CFG_HOME. See this post for more information. That part is important because it needs the configuration files in the new locations. Then, I deleted the Cache, the logs, and the log_output to clean things up. The app server still wouldn’t start, so I tried to reconfigure it. That’s when I received this error:
Enter selection (1-26, h, or q): 13 Warning: no changes detected, some configuration files will not be updated. Loading configuration... CMDTUX_CAT:111: ERROR: TUXCONFIG (C:\Documents and Settings\psadmin\psft\pt\8.51 \appserv\HR90DMO\PSTUXCFG) of machine SKPHCM90 must be the same as the TUXCONFIG environment variable (E:\psoft\HR90DMO\appserv\HR90DMO\PSTUXCFG) CMDTUX_CAT:867: ERROR: tmloadcf: Above errors found during syntax checking ==============ERROR!================ Unable to load the configuration! Check the TUXEDO log for details. ==============ERROR!================ ==============ERROR!================ Unable to load the Domain configuration! Check the TUXEDO log for details. ==============ERROR!================
This thread claims that you have to set the TUXCONFIG variable. I didn’t think that you did have to set it, so I did a little more experimenting.
In the config file, I found the Log/Output Directory was still set to the “Documents and Settings” directory:
I changed this to the PS Home directory:
So, that fixed it so that I could reconfigure the process scheduler via psadmin, but I still had trouble with the report distribution. Process Monitor said it was running without a report node:
I went to check the Report Node (PeopleTools > Process Scheduler > Report Node). The port number was wrong on the URL, so I had to fix that (it should probably match the port number on your browser as you are logged in):
That still didn’t fix the problem, so I checked the server definition (PeopleTools > Process Scheduler > Servers). Sure enough, the report node was blank!
So, I entered the report node and saved the page:
2 thoughts on “Troubleshooting Report Distribution”
I also experimented about the missing report node. It is emptied by the Peopletools upgrade scripts. Sure, we can spent quite some time to know why it is not posting…
And still, you did not have upgraded an existing PS_HOME but apparently created a brand new home. Otherwise you would have seen (and maybe error out) an other difference, regarding psadmin tool from %PS_HOME%appserv pointing to %PS_HOME%bin and you should do the change manually (find out more E-UPG PT 8.51 New location of Psadmin.exe [ID 1251444.1]).
I always appreciate your comments. They are always helpful.
I actually did upgrade an existing PS_HOME. I saw your post on your blog about the psadmin link. I am on Windows though, so I know it couldn’t do a symbolic link to it in the bin directory.
I have been trying to get my image clean and stable so I can snap a backup of it. Then, I will do some more playing around with it. I will definitely go back and do some more looking into psadmin.
Right now I am fighting with the the report repository. I have the report posting, but now I can’t see the reports via the report repository. It is giving me an error about the site name not being valid. I need to do some more research, probably on the local portal node.