Stephen's Oracle Blog

We have a NetApp FAS filer attached via Gigabit Ethernet to a Sun V440 running Oracle databases the files for which are stored on the filer The NetApp volume is mounted under NFS version 3 at $ORACLE_BASE/oradata. Under that directory are 7 directories .snapshot, AXEM, AXER, AXEM_backup, AXER_backup, AXEM_nobackup, AXER_nobackup. .snapshot is where the NetApp filer stores snapshots of the volume. The server runs two databases, AXEM and AXER. The AXEM and AXER directories hold the datafiles (and a small tempfile) for the appropriate database, *_backup holds the RMAN and cold backup sets for the appropriate databases. *_nobackup holds the bulk of the tempfiles (temporary tablespace, used for sorting) for the appropriate database (when we do cold backups we don't bother to backup the tempfiles). For AXEM, AXER and *_backup both du and ls report the same file sizes for both data and tempfiles. For *_nobackup ls reports the correct sizes (as specified in the Oracle data dictionary) but ...

I was just looking at my stat counter analysis for this blog, it's usually a good way to find out if anyone has linked to me. From the results it looks like the vast majority of people (around 97%) find this blog through a search engine, virtually all (over 99%) through Google. Most common search strings are or include " ORA-07445 ", " opmnctl failed to start process " and " ORA-01031 ".

During his presentation on " Inside RAC " Julian Dyke appeared to suffer laptop difficulties so asked for 4 volunteers to assist in a demonstration. Being the helpful soul (i.e. idiot) that I am, I volunteered. Mr Dyke did seem remarkably wellprepared in terms of props, almost as if he had been expecting his laptop to have problems :-)! Photographic evidence (courtesy of Connor McDonald) can be found here and here . I'm the fat, ugly, balding, bearded one on the far left, aka Node 1. Node 2 was Babbette Turner-Underwood , Node 3 (aka Master) was Murali Vallath and Node 4 was Carel-Jan Engel .

A question just appeared on the Oracle-L mailing list about whether there are any issues with using histograms. I've seen a number of presentations over the last couple of years that have warned about issues where you have histograms on highly skewed data and use bind variable (or have cursor_sharing=force, which synthesises bind variable from literals in queries), so I responded: On 19/11/06, A Joshi wrote: > Hi, > About use of histograms : I think histograms are useful for indexes on > columns that are not very selective. However I came across note 1031826.6 on > metalink. About maintenance and space cost. I think space cost is negligible > and can be ignored. About maintenance : does it mean statistics need to be > gather often? Or does it mean some other cost. > > Question : Is there any other overhead or any other negative impact of using > histograms? One downside I've been hearing a lot about over the last couple of years results from Bind ...

Today was the first day of the United Kingdom Oracle User Group 2006 conference . The day itself was pretty good, lots of good talks (more on that later), but I ran into a couple of issues in the evening. The first issue was when I went into the Handmake Hamburger just accross the canal from the ICC , I was alone and the waitress/Maitre d' said they didn't have any free tables suitable for one person (looking around they had several free tables, just they were for 2, 4 or more people) so after expressing my disatisfaction and feeling about the poor quality of service I went accross Broad Street to Walkabout and had a Kangaroo Burger. The second issues was that I had hoped to go to the bloggers' meetup at All Bar One. Although I blog about Oracle I'm not sure if I qualify as a true Oracle Blogger so when everyone disapeared upstairs to a private party (UKOUG had laid on some sort of 'Thank You' event for the volunteers) I didn't follow. I know I'...