DBA in the Trenches

There are lots of logs created in a system and like cleaning the house over the weekend you also need to keep the system clean. Sometimes this can free up quite a bit of space in the storage. crontab is my preferred method of doing this sort of scheduled cleanup although it can be niggly at times. # Minute  Hour    Day of Month   Month    Day of Week   Command # (0-59)  (0-23)  (1-31)         (1-12)   (0-6) 30   11   *   *   6   find /u01/archive/*.dbf -mtime +7 -exec rm {} \; So this means that I want the script to run once a week on a Saturday as 11:30 in the morning. It will find all the archives (*.dbf) that are older that 7 days and remove them. Really powerful so you also need to becareful how to use this command. Great information of the power of crontab

In R12.2 ETCC is used to check for bug fixes on the DB and Middle Tier. It is required to run the ETCC script on the database tier after each clone as adop check's this as one of its preliminary check's so you may as well keep this one up to date. Lastest version can be found at 17537119 . Prepare the files to run them mkdir /stage/install/ETCC chmod 777 /stage/install/ETCC cd /stage/install/ETCC unzip p17537119_R12_GENERIC.zip -d /stage/install/ETCC chmod a+x /stage/install/ETCC/* Run the script on each of the tier's that the services are running on: Database Tier: echo $CONTEXT_FILE | /stage/install/ETCC/checkMTpatch.sh  Middle Tier: echo $CONTEXT_FILE | /stage/install/ETCC/checkDBpatch.sh  Reference:  How To Run The 12.2 EBS Technology Code Level Checker (ETCC) ? (Doc ID 2008451.1) 12.2 E-Business Suite - How to identify patches for listed ETCC missing database and middle tier bugs (Doc ID 2057925.1)

Mount old file system in new server so that you can copy the files over. Replace $[OLD_APPLCSF] ,$[NEW_APPLCSF] and [PROD] with your local settings rsync -a $[OLD_APPLCSF]/log/ $[NEW_APPLCSF]/log/ rsync -a $[OLD_APPLCSF]/out/ $[NEW_APPLCSF]/out/ Need to update the concurrent request table that stores the log file and out file locations update APPLSYS.FND_CONCURRENT_REQUESTS fcr set fcr.LOGFILE_NAME = replace  (FILE_NAME,'$[OLD_APPLCSF]','$[NEW_APPLCSF]') ,fcr.LOGFILE_NODE_NAME= '[PROD]' ,fcr.OUTFILE_NAME = replace  (OUTFILE_NAME,'$[OLD_APPLCSF]','$[NEW_APPLCSF]') ,fcr.outFILE_NODE_NAME= '[PROD]'; For Bi Publisher there is also the additional file that is created and this needs to be updated in FND_CONC_REQ_OUTPUTS. Sneaky one that I later found. update APPLSYS.FND_CONC_REQ_OUTPUTS set FILE_NAME = replace  (FILE_NAME,'$[OLD_APPLCSF]','$[NEW_APPLCSF]') ,FILE_NODE_NAME = '[PROD]';

With R12.2 you now have two file systems (fs): Run and Patch. Both actually being self explanatory. The run fs is where the current instance is being run from ($RUN_BASE) and patch fs is what adop uses to apply patches to before cutting over to use as the run fs ($PATCH_BASE). So one of the issues possibly for developers is that upon opening a ssh session they will need to source the run files system or even for DBA work you normally use the run fs unless specific work is needed on the patch fs. So how too make it easier than sourcing EBSapps.env each time and then entering either Run or Patch.  This is what is needed: source /u01/oracle/EBSapps.env run Now you could call this ever time you log into the instance or you could add it to $HOME/.bash_profile This way you will not have to worry about it and the person logging in will be using the run fs where most work is needed. Reference: How To Automatically Set the Current Run or Patch Edition / File System for EBS 1