Posted by Sheeri Cabral on Jul 29, 2010
At Kscope this year, I attended a half day in-depth session entitled Data Warehousing Performance Best Practices, given by Maria Colgan of Oracle. In that session, there was a section on how to determine I/O throughput for a system, because in data warehousing I/O per second (iops) is less important than I/O throughput (how much actual data goes through, not just how many reads/writes).
The section contained an Oracle-specific in-database tool, and a standalone tool that can be used on many operating systems, regardless of whether or not a database exists:
Read the rest of this entry . . .
Posted by John Scoles on Dec 3, 2009
I finally had the opportunity the other day to try and install DBD::Oracle on an IBM AIX 5.1 box, and for once I have some good news to tell.
Anyone who has ever tried this will know of some of the troubles I speak of. When dealing with DBI and any DBD on a AIX box, you either must either be lucky enough to have the same compiler installed that built the version of Perl that comes with the box (I have never seen this happen); or you have to spend a great deal of time downloading and installing your own GCC and the building your own version of Perl.
Fortunately, all the hard work was done for me by other members of my team, and I was only a Johnny-come-lately to the whole process.
In our case, these are the steps that we followed:
- get a working version of GCC on the box
- rebuild and install your own version of Perl
- build and install DBI using your local version of Perl
All of the above worked without a major problem. It was only when we tried to build DBD::Oracle that we ran into a problem:
Read the rest of this entry . . .
Posted by Luke Davies on Sep 30, 2009
Recently I was looking into a long-running statement and noticed a curious thing. One moment, I had just over 4 billion buffer gets, and the next I had around 2 million.
The statement was still processing and the number of physical reads was still rising.
I think the reason for this is that the internal place holder for this value is simply a 32-bit unsigned integer that has the largest number of 4294967295. But my version of Oracle is 64-bit; I thought that maybe it would use 64-bit integers.
So, beware that if you have had a statement running for some time, you cannot necessarily rely on the buffer_gets column in v$sql—it may be that it has run over the limit, been recycled, and is counting from zero again.
The evidence. Read the rest of this entry . . .
Posted by Darrin Leboeuf on Nov 21, 2007
It recently came up that it would be helpful if we had a cheat sheet to find out the machine names for any given UNIX. I knew these off the top of my head but it would be great if people added more as comments.
HP/HP-UX: /bin/uname -i
IBM/AIX: /bin/uname -m
SGI/IRIX: /sbin/sysinfo -s
Sun/Solaris: /usr/ucb/hostid
Posted by Marc Fielding on Nov 12, 2007
The 11g platforms are now coming out fast and furious:
Windows 64-bit
Solaris SPARC 64-bit
AIX PPC64
HP-UX Itanium
In addition to the previously-released:
Linux x86 32-bit
Linux x86 64-bit
Windows 32-bit
So download away, after checking your platform certification first, of course.
Planned future platforms:
Apple Mac Intel OSX
HP-UX PA-RISC
OpenVMS Itanium 64-bit
Linux LPAR on IBM z-series mainframes
Linux on IBM Power
Linux on Itanium
Solaris x86 64-bit
