Here are some very good presentations from various Oracle conferences which I attended and which I thought I would share with the wider community.
I am still in the process of uploading ones which I think will be most useful, so this page will be updated on a regular basis as there are about 200 presentations to upload.
Please Note:
If anyone has objection to these presentations being made available for viewing by the general public or if it is in breach of any copyright restrictions, then kindly do let me know and I will remove the presentation from the web site.
Thanks to all the authors of these presentations for their good work.
The following two tabs change content below.
Gavin Soorma
Hi Gavin
When do we need RAC really?
Out of my own experience,I have seen RAC doing ‘bad’ at times.
Performance is not so good in RAC machines(lot of ‘gc’ wait events clinging up space and time )
Please advise
Regards,
Hi Shafi – there is no real right or wrong answer to this. RAC basically affords two advantages which is availability and scalability. In terms of availability if one node or server goes down, the application will continue to run on the other surviving node or nodes. Also, from the scalability point of view, RAC enables you to provide capacity on demand. So as the application user base grows, it may not always be possible to grow vertically as you may be hitting machine or OS limits in terms of adding CPU – you then need to grow or scale out horizontally by adding additional nodes to cater for increased capacity requirements. RAC is NOT the answer for all applications and needs to be carefully considered as it is a different kettle of fish as compared to a single instance database. However, as far as performing poorly goes I do not think that would be a general case for RAC – and in almost 99% of the cases, the problem always lies in the interconnect between the nodes in the cluster which is in most cases responsible for all the ‘gc’ related waits.
your answer is well appreciated,thanks!
Dear Gavin:
I am using ASM RAC box(10204) on Linux RHEL3(ocfs v2,perhaps) – suddenly,there are few messages saying “ORA- Diskgroup space exhausted” while trying to add datafiles….(weirdly,the DG +DATA1 has ample amt. of freespace on it – >asmcmd df -k )
I looked into the matter by myself and studied that there are some issues with DG rebalancing…Per some metalink notes etc,I used “alter DG rebabalance power limit 11” before I successfully added few datafiles…
I emperically used this command twice or thrice and observed the progress of this op – v$asm_operation and >asmcmd df -k
I understand amending Power limit(at the memory) for disk – rebalancing to higher value than default 1 etc,will hamper the performance of your whole system with more and more CPU cycles……that too being a 2tb VLDB system used for Compliance…I don’t need to take a chance either…
What is your suggestions?
Regards