I would like to ask you a question regarding a previous question ("ORA-1652: unable to extend temp segment by 1280 in tablespace TEMP") posted on your site. In reply to this question you say ". . . Oracle 8i and 9i when using Locally Managed Tablespaces (LMTs) for TEMP tablespaces when the applications maintain persistent connections, that the instance does not release unused temporary segments for future use by another process."
- With regard to this bug what version was this behavior fixed in?
- How would not releasing temp segments affect future connections?
My reason for asking is that we are on 9i SE v22.214.171.124 for Solaris, and without warning or being able to replicate at will we have two issues that could be caused by the same underlying issue: a) ORA-100: no data found, where data is clearly in the schema, and b) sometimes I see data written to the incorrect schema (were we have more than one schema which is almost the same on the same instance (develoment box)).
We cannot replicate this behavior at will, but if I turn on session tracing or flush shared_pool the problem goes away! Today I noticed a couple of occurances of the ORA-1652 error for both TEMP and other instance tablespaces in my alert.log over the last few weeks. Could these be linked?
I would appreciate whatever feedback you may have to offer.