This content has been marked as final. Show 3 replies
You mentioned you're in NOARCHIVELOG mode but you're getting an ORA-00308 error. This might be happening if the LOG_ARCHIVE_DEST_n parameter is still pointing to an archive log destination and if the default value for the LOG_ARCHIVE_DEST_STATE_n parameter is set to 'ENABLE'. In this case the database database will still attempt to open the required archive log file during recovery of a long running transaction.
To resolve that issue you could set LOG_ARCHIVE_DEST_STATE_n to 'defer'. See: http://docs.oracle.com/cd/E11882_01/server.112/e25513/initparams124.htm#REFRN10087
However that issue is really just a side effect of having to rollback a long running transaction, so the main problem still exists.
I checked MOS for any mention of mdtnmconnect() or mdtngt_edge. I found two old reports...
Bug 7176267 : ACCESS_VIOLATION WHILE BUILDING A TIN, DEPENDING ON BLOCK SIZE USED
Bug 6606678 : CREATION OF A TIN FAILS WITH ORA-7445 [MDTNMMKEDGE]
... but there doesn't seem to be any conclusion in either of those. If anything it sounds like it might be worth playing around with the blk_capacity parameter to see if it helps (i.e. trying increasing and reducing it).
Have you checked through the kugioas_ora_58460.trc trace file for anything of potential use?
Apart from some basic prototyping I haven't really used SDO_TIN so can't comment much on how to troubleshoot it.
By the way, what's your exact Oracle version - just the base 18.104.22.168 or have you added any patch bundles (if not, then you probably should) as they might help.
Edited by: John O'Toole on Apr 30, 2013 9:02 AM
First of all, thanks for the response.
1. Thank you for your clarification about LOG_ARCHIVE_DEST_STATE_n parameter. I didn't know about it.
2. I tried to use blk_capacity parameter=5000, 50000, 100000. Unfortunately, it doesn't help.
3. kugioas_ora_58460.trc trace file doesn't contain any additional useful information:
Trace file C:\ORACLE\diag\rdbms\kugioas\kugioas\trace\kugioas_ora_58460.trc
Oracle Database 11g Enterprise Edition Release 22.214.171.124.0 - 64bit Production
With the Partitioning option
Windows NT Version V6.1 Service Pack 1
CPU : 24 - type 8664, 12 Physical Cores
Process Affinity : 0x0x0000000000000000
Memory (Avail/Total): Ph:11725M/32757M, Ph+PgF:45823M/65513M
Instance name: kugioas
Redo thread mounted by this instance: 1
Oracle process number: 94
Windows thread id: 58460, image: ORACLE.EXE (SHAD)
*** 2013-04-30 01:28:12.989
*** SESSION ID:(447.1) 2013-04-30 01:28:12.989
*** CLIENT ID:() 2013-04-30 01:28:12.989
*** SERVICE NAME:(KUGIOAS) 2013-04-30 01:28:12.989
*** MODULE NAME:(TOAD 126.96.36.199) 2013-04-30 01:28:12.989
*** ACTION NAME:() 2013-04-30 01:28:12.989
Exception [type: ACCESS_VIOLATION, UNABLE_TO_WRITE] [ADDR:0x44] [PC:0x146A7E2F0, mdtnmconnect()+76]
Incident 55960 created, dump file: C:\ORACLE\diag\rdbms\kugioas\kugioas\incident\incdir_55960\kugioas_ora_58460_i55960.trc
ORA-07445: exception encountered: core dump [mdtnmconnect()+76] [ACCESS_VIOLATION] [ADDR:0x44] [PC:0x146A7E2F0] [UNABLE_TO_WRITE] 
ORA-13199: Error in mdtngt_edge
4. I'll try to install patches on database. It looks promising.