Maybe PMON failed to hand the rollback of killed session's transaction over to SMON properly, maybe got in some sort of loop. And being in this loop it couldn't move on to clean up the v$process entries. But yes, I've been also in similar situation - where the decision to bounce the instance has been done quite quickly. Better to have whole instance up in determined amount of time than having the instance in partially working state for undefined period of time (and perhaps having to bounce it eventually anyway). Tanel. > -----Original Message----- > From: oracle-l-bounce@xxxxxxxxxxxxx > [mailto:oracle-l-bounce@xxxxxxxxxxxxx] > Sent: Thursday, January 11, 2007 01:41 > To: Tanel Poder; oracle-l > Subject: RE: Rollback waiting on 'cache buffer chains' latch - SOLVED > > No, the physical process was gone (killed with kill -9): > > 15:34:17 SQL> select b.spid from v$session a, v$process b > where a.paddr=b.addr and a.sid=664; -- //www.freelists.org/webpage/oracle-l