I don't think so Insert/Merge would have generated this DDL. The best way is to work with oracle support for this kind of errors(600/7445). I do see some bugs related to memory/data corruption for 600 errors but without more information can't say anything On Mon, Apr 8, 2013 at 11:14 AM, <Christopher.Taylor2@xxxxxxxxxxxx> wrote: > I've just begun troubleshooting some various Oracle issues with invalid > ROWIDs, ORA-600 [25027], ORA-600 [kcbzfb_1], ORA-08103: object no longer > exists errors. > The trace file dumped for the 25027 indicates an INSERT /*+ APPEND */ > statement and I'm running an analyze table validate structure for all the > tables/indexes involved in in the INSERT statement. > > While that is running, I cross-referenced DBA_OBJECTS LAST_DDL_TIME and > noted that many objects (some of which are involved in this query) are > showing INDEX DDL operations on partitioned indexes during the time frame > of the errors. > > For a direct-load INSERT using the append hint, would that account for the > INDEX DDL operations I'm seeing? (I should probably just do a test to > confirm) but wanted to poll the list first. (I'll do a test anyway while > I'm waiting for a response while I'm thinking about it). > > We use many partitioned tables and we have 50-100 concurrent sessions > processing distinct ORG_IDs that should be in their own partitions but I do > see a PK INDEX showing DDL operations on one of the tables involved in the > INSERT statement (the PK index is on a table involved in the SELECT portion > of the INSERT). > > Regards, > > Chris Taylor > Oracle DBA > Parallon IT&S > > > -- > //www.freelists.org/webpage/oracle-l > > > -- Thanks & Regards, Taral Desai -- //www.freelists.org/webpage/oracle-l