Posts Categorized: Oracle
15 years ago, with the release of Oracle 7.0.12, Oracle gave the world—or at least its customers—something really great: the Oracle Wait Interface (OWI). We believe that the Oracle database software is the best instrumented database software available. The fact that Oracle already leads the industry in this regard probably led to their decision to make this leap forward in instrumentation an extra-cost item. However, in the interest of making Oracle even better, we would like to invite readers to join us in signing the following open letter to Larry Ellison, CEO of Oracle Corporation.
Before you proceed with reading this post, I strongly encourage you to read Tom Kyte’s trilogy about write consistency, since I’ll do only a brief introduction to the subject. The way Oracle ensures UPDATE write consistency is through a mechanism called restart. Let’s take a look at an example before we proceed with the main topic of this blog post, Will there be any difference if we substitute the following MERGE for the last UPDATE?
A few month ago I posted about my indignation regarding the inability to change my email address on OTN. Now, I’m not only able to change my email address, but also the screen name (I don’t think I do that before either). In the end, it took Oracle just 2 months and 5 days to follow up on my post. Not too bad, considering that OTN forums were full of complaints for years! ;-)
I’m looking forward to the announcement of the next Oracle database version. It’ll be interesting to see how it’s going to be named. It’s been known so far as 11g but I expect an update on 11th of July.
Let’s take the following question, for example. Is there any difference between using: where column between n and m and where column>=n and column<=m? Looks like a simple one, eh? hey are the same from a semantic point of view. But SQL is a declarative language. In other words, you wouldn’t expect same execution plan with two semantically identical statements, would you? There is at least one known (to me) example where both statement produce different execution plans. You never know until you test it. We start by creating a simple list-partitioned table with the local index:
My team and I still use old-style rollback segments for one of my client’s 10g production databases. We just never found the need to switch to automatic undo management. There are a number of 1GB rollback segments. They are that size because they need to be able to support large transactions. At the same time, we don’t want to have transactions bigger than 1GB as this is an OLTP system. For the past few weeks we’ve had a strange problem. One of the web calls would cause one of the rollback segments to become full by using 1GB of undo data.
Oracle 10g introduced the ability to run shared server sessions within the workarea_size_policy=auto parameter. However, there is one caveat I would like to point out. I noticed this after one of our databases was upgraded from the 9iR2 release.
Here are a few nice typos that had quite disastrous consequences. After having fought some network problems to get a distribution of Oracle installation binaries from OTN to a Linux box, a colleague found a revolutionary way to unpack a cpio archive. Have a look.
Since Oracle 10.2 release is getting more and more popular in production settings, It’s especially useful to look on it vis-a-vis upgrades. In addition, there is a small gotcha that might cause performance problems in certain extreme cases.
I’ve never really liked the idea of REMOTE_LOGIN_PASSWORDFILE=SHARED, probably just because I haven’t seen much use for it. As a result, I’ve never paid any attention to it. If you don’t recall the difference between EXCLUSIVE and SHARED settings for 9i, here is the quote from the documentation.