11g Release 2

martin.bach's picture

Performance testing with Virident PCIe SCM

Thanks to the kind introduction from Kevin Closson I was given the opportunity to benchmark the Virident PCIe flash cards. I have written a little review of the testing conducted, mainly using SLOB. To my great surprise Virident gave me access to a Westmere-EP system running a top of the line 2s12c24t system with lots of memory.

In summary the testing shows that the “flash revolution” is happening, and that there are lots of vendors out there building solutions for HPC and Oracle database workloads alike. Have a look at the attached PDF for the full story if you are interested. When looking at the numbers please bear in mind it was a two socket system! I’m confident the server could not max out the cards.

Full article:

Virident testing martin bach consulting

martin.bach's picture

Tale of a platform migration Solaris 10 SPARC to Linux

This is as much a note to myself how to do this in the future as it is something hopefully worth reading for you. The requirement has been precise as always: migrate a database from 10.2 on SPARC to 11.2 on Linux. In the process, go from Veritas to ASM and make it quick!

I like short briefings but this was too short. Since the database was reasonably large I opted for the transportable tablespace approach, however I now think that a massively parallel impdp with network_link could have saved me quite a bit of time.

The following is by no means  the complete story, but hopefully gives you an idea how to do these things. Always check, and document, then test (rinse and repeat). Only when proper signoff is received should you try such a process in production. Remember to script it and have at least one clean run of the scripts! This process is not super-quick, if you have low downtime requirements then consider Streams or better: Golden Gate for the process.

martin.bach's picture

Little things I didn’t know: difference between _enable_NUMA_support and numactl

In preparation for a research project and potential UKOUG conference papers I am researching the effect of NUMA on x86 systems.

NUMA is one of the key features to understand in modern computer organisation, and I recommend reading “Computer Architecture, Fifth Edition: A Quantitative Approach” from Hennessy and Patterson (make sure you grab the 5th edition). Read the chapter about cache optimisation and also the appendix about the memory hierarchy!

Now why should you know NUMA? First of all there is an increasing number of multi-socket systems. AMD has pioneered the move to a lot of cores, but Intel is not far behind. Although AMD is currently leading in the number of cores (“modules”) on a die, Intel doesn’t need to: the Sandy-Bridge EP processors are way more powerful on a one-to-one comparison than anything AMD has at the moment.

martin.bach's picture

RMAN restore spfile from autobackup and compatible weirdness

Prompted by a recent interview I wanted to perform a little test with RMAN, and incrementally updated backups. I created a database on my Linux (OEL 5.5) test system and refreshed my understanding of this most useful rman feature (which is for another post).

So after I was happy with the working of the incrementally updated image copies I decided to see if I could restore my database with those. Time to use “drop database”, which removes spfile, data files, temp files and the control files. Tabula rasa!

But it didn’t matter I thought: I have controlfile/spfile autobackups and fully recovered image copies plus all the archived logs in the FRA. What could possibly go wrong? Well it took me 30 minutes to get the database back.


martin.bach's picture

WARNING: FAILED to load library: /opt/oracle/extapi/64/asm/orcl/1/libasm.so

This post is long overdue, and it’s as much a reminder to myself as it is a post for you to find when scratching your head.

Here’s the situation: Oracle Restart on an Oracle Enterprise Linux 5.5 system with ASMLib in use. I have installed plus the latest PSU to bring database and Grid Infrastructure to far so good. I have used the non-GUI approach and everything seemed ok until I tried to create a database for some benchmarking. Here’s the surprising output:

martin.bach's picture

RMAN duplicate and read-only tablespaces

This one is to be filed under the “little things I didn’t really know until now” section: RMAN duplicate. Be honest to yourself: would you have known the answer to this question: what happens to tablespaces which are read-only on the source during an RMAN duplication (not for standby)?

Some background

I have started my career as a DBA on Oracle, and at the time not everyone fully embraced RMAN. OK, RMAN made it really hard at the time to fall in love with it. So when we “cloned” our production database to development, the following steps were followed:

martin.bach's picture

RAC Attack at OUGN 2012

Quick post to bring the Virtual Box PDF to the attendees. Follow this link to get the instructions

RAC Attack Using VirtualBox


This post was the best way to get the PDF instructions to participants of the OUGN 2012 RAC Attack seminar without too much effort on our side. If you have since then used your favourite search engine and ended up here in the hope that there is an interesting story please navigate here instead:


Thanks for passing by!

martin.bach's picture

No more excuse not to use large pages in Linux

This is a very quick post to remind users of the importance of large pages. Before it was a little more elaborate process to determine the number of large pages to be set aside for Oracle, but now it’s so simple there isn’t any more excuse not to use large pages. If you have > 16G SGA and lots of processes, check the impact of large pages in /proc/meminfo before and after. By implementing large pages at a former company I reduced the page tables overhead from 20G (60G sga, processes set to 1000, most often around 800 sessions connected) to a fraction of that number.

Here is why you can’t hide anymore:

martin.bach's picture

Oracle client not relinking properly

One of the tasks I am performing quite regularly is to deploy Oracle software in form of an RPM. In a previous post I described how this proces could work, based on a post by Frits Hoogland.

Employing the same method, I ran into problems with Oracle 11.2.0.x clients. A few facts to start with:

  • Oracle client 64bit
  • Golden image created on Oracel Linux 5
  • Destination: SuSE Enterprise 10 SP4

The problem described here is most likely applicable to other Oracle clients as well although I haven’t verified that.

The problem

martin.bach's picture

Did you know the cluvfy healthcheck?

While I was performing a three day seminar recently in Switzerland I came across this new option in cluvfy.

Normally you’d run cluvfy in preparation of the installation of Grid Infrastructure or a set of RAC binaries to ensure everything is ready for the next step in the RAC install process. Beginning with, there is another option that’s been sneaked in without too much advertisement: the healthcheck.

Part of the “comp” checks, it takes the following options:

cluvfy comp healthcheck [-collect {cluster|database}] [-db db_unique_name] [-bestpractice|-mandatory] [-deviations] [-html] [-save [-savedir directory_path]

The most extensive report is run without any options, as shown in the appendix (the output is too long to display at this stage of the post) You have the following options:

To prevent automated spam submissions leave this field empty.
Syndicate content