Batch Server Reboot, Cache Cleared and IPC Cleanup Prevents Fatal SQL Errors?

We have a situation where if we don't reboot, clean cache and IPC on a weekly basis for our Linux batch servers, then we start to see many fatal SQL errors and PeopleCode RC = 24  . . . .they usually reference different SQL errors and all seem to go away when the server cache has been cleared and rebooted . . . anyone else seen this?   We are on PS Financials 9.2 missing many of the patches since the original release , PeopleTools 8.53

Here are some examples:

2:32:55AM  Process is blocked because maximum concurrent instance (1) is reached   

   2:35:41AM  1 File(s) to process.   

   2:35:41AM  File: /vob/peopletools/src/pspcm/pcmobuf.cppSQL error. Stmt #: 5574  Error Position: 0  Return: 1001 - ORA-01001: invalid cursor Failed SQL stmt: INSERT INTO PS_TR_EVENT_LOG (EVENT_TRACKING_ID,EVENT_CODE,EVENT_SEQUENCE,EVENT_DATE ,EVENT_SEVERITY,EVENT_PROCESSED,OPRID,MESSAGE_SET_NBR,MESSAGE_NBR,EVENT_MSG_BIND1,EVENT_MSG_BIND2,EV ENT_MSG_BIND3,EVENT_CHANGED_SW) VALUES (:1,:2,:3,TO_TIMESTAMP(:4,'YYYY-MM-DD-HH24.MI.SS.FF'),:5,:6,: 7,:8,:9,:10,:11,:12,:13)