hanganalyze doesn't show any issues ... it was a lock type 'CU' => cursor bind. I deduced (probably correctly) that it could be a _feature_ of cs=force. so much for cs=force ... Raj On Wed, 23 Mar 2005 18:57:48 -0500, Paul Drake <bdbafh@xxxxxxxxx> wrote: > try hanganalyze instead. > less filling. > > finally found my first "unable to extend segment in <undotablespace> > by n bytes" message in 9.2 today. That was a pretty good run without > one. > > Paul > > > On Wed, 23 Mar 2005 15:52:14 -0500, rjamya <rjamya@xxxxxxxxx> wrote: > > for lock type 'CU' what do values id1 and id2 indicate? > > > > We had a developer experience DBDS (death by dynamic sql). The session > > is waiting on 'global cache cr request' waiting for more than 3000 > > seconds, 10046 trace file doesn't get updated, and v$lock shows type > > 'CU'. > > > > BTW using oradebug to dump processstate for the process in question > > doesn't work, I had to kill the dump processstate command after 4 > > minutes. So, I dumped the systemstate and found some more information. > > > > So once again the question is what do the values of id1 and id2 > > indicate for lock type 'CU'. > > > > After looking 'CU' I had developer put in 'cs=exact' in the code and > > the report finished in expected time (as against 5 hours and > > counting). > > > > I am just curious I guess. > > TIA > > Raj > > ------------------------------ > > select standard_disclaimer from company_requirements where category = > > 'MANDATORY'; > > -- > > //www.freelists.org/webpage/oracle-l > > > > > -- > #/etc/init.d/init.cssd stop > # f=ma, divide by 1, convert to moles. > -- ------------------------------ select standard_disclaimer from company_requirements where category = 'MANDATORY'; -- //www.freelists.org/webpage/oracle-l