Long time since I blogged, but I have been busy with tinkering with Ubuntu, learning about Oracle streams, client projects and playing with my now 10 month old son (not necessarily in this order).
Still, to let you know that I am still alive: a short overview of things I learned that you might find useful:
Tanel Poder’s advanced troubleshooting course
If you still have the possibility to follow it, then do so. I really learned a lot during this course.
The word “advanced” in the course title is not soley there for marketing purposes
More information about the course contents can be found here
Datapump and lobs
Until 10.2.0.4 there is a bug in datampump with lobs, causing them to be slow on importing. This is due to bug 5555463, which is fixed in 10.2.0.4 (there is also an one-off patch available). Another work-arround is to enforce the access_method used by datapump, but this is unsupported.
Some articles about this that I found usefull:
Note:5555463.8 – Bug 5555463 – Slow performance of datapump IMPORT with small LOBs
Note:552424.1 – Export/Import DataPump Parameter ACCESS_METHOD – How to Enforce a Method of Loading and Unloading Data ?
Note:286496.1 – Export/Import DataPump Parameter TRACE – How to Diagnose Oracle Data Pump
Netapp Snapmanager for Oracle 3.0 has been released.
This release has some important new features, such as time based retention policies.
Don’t have the time to play with it myself, but one of my colleagues has started to test it, and is blogging about it: http://bjornnaessens.wordpress.com
Autocompletion for hostnames
Prevent windows from requiring to reactivate after a P2V
I created a vmware image from my laptop to run it under Ubuntu, unfortunately my laptop was installed with an OEM edition of XP, so it kept comlaining that I needed to reactivate the license again.
Luckally there is a vmware parameter you can use to avoid this: http://communities.vmware.com/thread/99057
(Thanks to Filiep Declerck for the link)
10.2.0.4, opatch, solaris and rac
With 10.2.0.4, opatch does not seems to update the $ORACLE_HOME/inventory/oneoffs/ and the $ORACLE_HOME/.patch_storage/ directories on the remote nodes when applying one-off patches. We have noticed this issue on several servers with Solaris 10 and oracle 10.2.0.4 (and opatch 10.2.0.4.3 and 10.2.0.4.5). Other combination might have the same problem.
The patch itself seems to be applied on the remote nodes, but because the directories are not copied a query (such as lsinventory) on the remote inventory will fail on the remote nodes (preventing installing other patches).
This issue is now tracked in bug 7611827. The workarround is to use the “-local” flag and install the patch node per node.
That’s all for now.