I am not a DBA, but that doesn’t mean I lack respect for how difficult planned upgrades can be in complex Enterprise deployments. If I was a DBA planning an upgrade of my Oracle Database 10g Release 2 database I would be looking forward to 10.2.0.4. From what I’ve seen it looks like a pretty substantial release. Metalink 401436.1 spells out some of the bugs that could/should be fixed in the 10.2.0.4 release.
Those of you who have read my NUMA-related posts might be interested to know that Metalink 401436.1 lists bug 5173642 as once of the bugs that could/should be fixed in 10.2.0.4. That bug was essentially a deadline on NUMA optimizations in 10gR2 as is clear by the specified workaround:
Workaround: Do not use NUMA optimization. eg: Set the following init.ora parameters: _enable_numa_optimization=FALSE
_db_block_numa=1
Kevin Closson's Blog: Platforms, Databases and Storage 
I am in the process of setting up a new environment for a new initiative using 11.5.10 and 10.2.0.3. It is too bad that 10.2.0.4 will come out probably in January and then get certified with 11i after a few months.
Hello Amir,
Yes, unfortunate timing that it isn’t coinciding with year-end planned maintenance (10.2.0.4 that is)… BTW, someone from your management chain stopped by to say hello after I spoke at OOW 2007.
Hi Kevin,
The person who spoke with you was my manager, Chuck Schlegel, and I had asked him exclusively to find time and attend to your session. He was very impressed. Now that you are with Oracle, we may be able to arrange a phone conversation with you to discuss our future direction with Oracle technology, specially Linux, NFS/NAS and clustering.
I have just finished reading all of your Opteron/NUMA post. I know you are a big fan of AMD and the HP DL585. Have you done any comparisons of the HP DL585 with an HP DL580? Is the DL580 a NUMA machine? Which one would you by today for a RAC cluster?
Hi
I have these Problems with my database
ORA-16038: log 3 sequence# 226 cannot be archived
ORA-19809: limit exceeded for recovery files
ORA-00312: online log 3 thread 1:
‘ /u10/app/oracle/oradata/PAEC/onlinelog/o1_mf_3_45bh430j_.log’
ORA-00312 online log 3 thread 1:
‘ /home/oracle/flash_recovery_area/PAEC/onlinelog/o1_mf_3_45bh45w5_.log’
It is worth noting that the problem has not been fixed in 10.2.0.4 see bug 7171446.
We are using 10.2.0.3 and not facing any of the BUG mentioned. But our OS & hardware is NUMA
enabled. If we disable OS (Linux x86_64) & Hardware (AMD Optron Server on HP Proliant boxes)
level then performance decrease significantly.
The best performance we got is with NUMA enabled at every level i.e. Hardware/OS/DB (by
_enable_numa_optimization=TRUE). We are not changing any other NUMA init.ora parameter e.g.
_db_block_numa. We have 4 NUMA nodes so do we need to set this _db_block_numa=4?
numactl –hardware
available: 4 nodes (0-3)
node 0 size: 8056 MB
node 0 free: 651 MB
node 1 size: 8080 MB
node 1 free: 1656 MB
node 2 size: 8080 MB
node 2 free: 1759 MB
node 3 size: 8080 MB
node 3 free: 1694 MB
node distances:
node 0 1 2 3
0: 10 20 20 20
1: 20 10 20 20
2: 20 20 10 20
3: 20 20 20 10
Also even NUMA enabled at DB level, SGA (24 GB) distribution is not happening evenly on NUMA
nodes. Do we need to upgrade the DB patchset to 10.2.0.4 and apply the Patch 8199533 for
effective NUMA usage?
following is the NUMA stat:
>numastat
node0 node1 node2
node3
numa_hit 321351567 42297617 61060565 42997951
numa_miss 700232 23301797 37434003 29815244
numa_foreign 86695735 3412301 844487 298753
interleave_hit 7082 25292 28909 29702
local_node 321347157 42236010 60981069 42920729
other_node 704642 23363404 37513499 29892466
Clearly, Node0 is more burdened with memory requests.All the Oracle & other internet forum discussion suggest to upgrade to 10.2.0.4 and apply DB Patch 8199533 for efficient NUMA utilization. We cannot afford to disable NUMA at OS & hardware.
I’m sorry to say I have not touched a 10g instance in over 4 years.