temp segments

I have a question about how oracle version 7.0.16 is managing space. While cleaning up a large amount of data I noticed the following. We had approx. 3000 tables spread across 10 tablespaces. I dropped each table prior to dropping the individual tablespaces(I realize I could have used the including contents, however in the past that has required upping initora parms to very high numbers to do this kind of cleanup). Anyway, after I delete 'ALL' the objects from these tablespaces, I drop the tablespaces. When I did the drop, I did not specify the including contents (just to make sure they really were empty). To my surprise oracle said there were objects still in 6 of the 10 areas. I looked through all of the dba_???? views trying to find what was in these areas. All views reported nothing. I finally looked into the sys.seg$ table and saw what appeared to be clusters in each of my tablespaces. I queried and saw these in each of my tablespaces. After about 30 minutes, I tried to query the seg$ looking for the 'clusters' in one of the tablespaces. No rows were returned for any of the tablespaces that 30 minutes prior had clusters in them. I then could issue the command 'drop tablespace xxxx' and while it previously failed, it now worked. I did not remove a single object between the initial drop tablespace failing and the subsequent drop tablespace succeeding. Is this somehow related to the smon process that is responsible for cleaning up temp areas. One additional note, I see very high amounts of cpu time being used by the smon process for the database that experienced this behavior. Please e-mail and I will summarize. Thanks for the time

Damian Yenzi (dyenzi_at_melpar.esys.com) E-Systems
Sys. Admin/Dba Received on Mon Aug 15 1994 - 23:49:14 CEST