I'm trying to keep this as short as possible, so please bear with me. I'm looking for any guidance on my 3.2 to 3.4 upgrade.
I'm working on a test server for Oct 8th upgrade from 3.2 to 3.4. Actual installation wasn't too bad. I'm using the "manual" upgrade method where I import the 3.2 dump, then run the ilink patches. The upgrade is fast approaching, yet I am still struggling with perforamance issues. My test server is a dual xeon 500 mhz with 2GB ram, which is barley meeting minimum requirements for a database sized as "Large".
PRO\LOCATE: Was slow, solved by TAN 128919, Set cursor_sharing = EXACT scope = both
CHECK OUT LARGE ASSY: Include generics, Include Instances, Include Drawings, all relationships. On our current server, this gathering process takes about 15 minutes and returns about 6000 parts. The test server takes over 2 hours. This same OLD test server with 3.3 took less than our current production server for the 3.3 test.
I am trying different CBO settings, described in TPI 110929. When I set optimizer mode to RULE and ran ...analyze _schema('PDM','DELETE'), my checkouts returned the "contact tech support". Support had me recompute the scema. This seems to go against the cancelling CBO logic. I've tried various settings here and its all very slow.
FILE TRANSFER ERRORS: When I do get a list of objects for a large assy to the check out screen, I get some file transfer errors. It seems 3.4 is finding different relationships than previous versions. The file it says it cannot find (encrypted name) does not exist in the vault on my test server or production server. And when I do a missing objects report, these parts do not show up as missing at all. Why is it trying to pull out non-existent versions?
Or is this all because my test server is outdated. I'm a silver level support customer, so I will not have ptc support that weekend.