I’m ashamed to say, I’ve been slacking off in the blogging department. I believe, however, that I can plead attenuating circumstances considering that I’m busy doing terrible things to Dancer 2 and that I just dusted off the elf bonnet and began to churn out proposals for the Perl Advent Calendar.
But still, I felt like I should resurface for a few minutes, if only to give a token sign of life. So here goes.
This post is the first in a series of discussions … perhaps you have never had the pleasure of needing to change a character set. If you have not and soon do, purchase a good sleeping bag as you may be at the office late for a few days to get the task accomplished.
I made this blog title intentionally provocative. However it isn’t far from the truth. Don’t believe me? Continue reading :)
Any database upgrade suppose to change SQLs’ execution plans for the better. In 99 cases out of 100 this is exactly what happens. What to do with rest 1% of the SQLs? This blog post is exactly about such unlucky case I have resolved today. Hope that you will learn something out of it. As always let me know what you think using the comments section at the end of the post.
I find myself forgetting on how to disable Oracle AUTO_TASKs on regular basis. Therefore I have decided to document it under this small blog post. As an additional bonus I may hear others experience in the area. So if you disagree or have some other experience please do not hesitate to comment under this post.
Just received a page about “Corrupt block”. From the first look it kind of scary to receive such page 40 minutes before your shift ends. Right?
After examining an alert.log (see below) and looking through knowledge bases found an explanation that it isn’t as bad as I thought. I liked the explanation I found therefore I decided to share it with rest of the world. Here you go folks
In this blog post I am publishing my SLOB tests results and conclusions testing Oracle database IO performance placing data files on NFS directly and on ASM disk group located on NFS.
Timezones can be darn confusing. So I thought, wouldn’t be nice to be able to switch the times back and forth on a webpage, such that you don’t have to juggle the time differences in your head, but rather just see the full thing first from your timezone perspective, then from the other guy’s?
The trick is simple: bundle all the files to be shared into a tarball called shared-files.tar.gz. As there is now only that one file, which name always remains the same, any new install is conveniently clobbering the old version.
Like many good stories, this one also started with an innocent question from customer: “I often check “SELECT COUNT(*) FROM TABLE;” on our old 10.2.0.4 cluster and the new Exadata. The Exadata is slower than our old cluster by few minutes. Exadata is not good at count(*) ? or other problem?” Here’s how I fixed it.