performance

randolf.geist's picture

Single Value Column Frequency Histogram Oracle 12c and later

It is hopefully in the meantime well known that Oracle has introduced in version 11g a new algorithm to gather statistics on a table that no longer requires sorting for determining the critical Number Of Distinct Values (NDV) figure - it instead uses a clever "approximate NDV" algorithm which always reads 100% of the table data and therefore in principle generates very accurate statistics. This new algorithm gets used only when the ESTIMATE_PERCENT parameter to the DBMS_STATS.GATHER*STATS calls is left at default or explicitly passed as "DBMS_STATS.AUTO_SAMPLE_SIZE". This new algorithm is also required in case other new features like "Incremental Statistics" should be used.

In 12c Oracle improved this algorithm allowing the generation of Frequency and the new Top Frequency histogram types in a single pass. The new Hybrid histogram type still requires a separate pass.

fritshoogland's picture

Where does the log writer spend its time on?

The Oracle database log writer is the process that fundamentally influences database change performance. Under normal circumstances the log writer must persist the changes made to the blocks before the actual change is committed. Therefore, it’s vitally important to understand what the log writer is exactly doing. This is widely known by the Oracle database community.

The traditional method for looking at log writer performance is looking at the wait event ‘log file parallel write’ and the CPU time, and comparing that to the ‘log file sync’ alias “commit wait time”. If ‘log file parallel write’ and ‘log file sync’ roughly match, a commit is waiting on the log writer IO latency, if it isn’t then it’s unclear, and things get vague.

fritshoogland's picture

Oracle wait event ‘log file parallel write’ change

This post is about a change in how the time is measured for the event ‘log file parallel write’. This is important for the performance tuning of any change activity in an Oracle database, because with the default commit settings, a foreground session that commits changes waits in the wait event ‘log file sync’, which is a wait on logwriter activity, for which the wait event ‘log file parallel write’ always has been the indicator of the time spend on IO.

Log file sync
First things first: a foreground session normally waits on the wait event ‘log file sync’ when it commits waiting for its change vectors to be written to the online redologfile(s) by the logwriter. It is wrong to always assume a ‘log file sync’ will be present. If, somehow, the logwriter manages to increase the ON DISK SCN to or beyond the foreground session’s commit SCN, there will be no ‘log file sync’ wait event.

Jonathan Lewis's picture

IOT Bug

Here’s a worrying bug that showed up a couple of days ago on the Oracle-L mailing list. It’s a problem that I’ve tested against 12.2.0.1 and 19.3.0.0 – it may be present on earlier versions of Oracle. One of the nastiest things about it is that you might not notice it until you get an “out of space” error from the operating system. You won’t get any wrong results from it, but it may well be adding an undesirable performance overhead.

Jonathan Lewis's picture

Temp space

A question about hunting down the source of the error “ORA-01652 unable to extend temp segment by NNN in tablespace XXX” shows up on the Oracle-L mailing list or the Oracle developer community forum from time to time. In most cases the tablespace referenced is the temporary tablespace, which means the session reporting the error was probably trying to allocate some space for sorting, or doing a hash join, or instantiating a GTT (global temporary table) or a CTE (common table expression / “with” subquery). The difficulty in cases like this is that the session reporting the error might be the victim of some other session’s greed – so looking at what the session was doing won’t necessarily point you to the real problem.

connor_mc_d's picture

Exadata storage indexes

We had a question on AskTOM inquiring about how to handle the issue of only 8 storage indexes being possible on an Exadata engineered system. If you are unfamiliar with what a storage index is, they are part of the suite of features often referred to as the “secret sauce” that can improve query performance on Exadata systems by holding more metadata about the data that is stored on disk. You can get an introduction to the concept of storage indexes here.

connor_mc_d's picture

The holistic SQL tuning series

I did a set of articles for Oracle Magazine on a more holistic view of SQL tuning. What do I mean by “holistic”? It was a reflection of a common problem that I see when questions come into AskTOM, or when people in the community approach me at conferences, namely, there is an inclination to dive straight into the deepest levels of the tuning exercise:

  • “What index should I create?”
  • “Should I increase the parallel degree?”

etc etc. And as technical practitioners, it is an easy trap to fall into. We too often fail to step back and approach the problem from its true requirement – that of, satisfying a business need. We might end up deep in the code, but we should probably not start there.

Jonathan Lewis's picture

v$session

Here’s an odd, and unpleasant, detail about querying v$session in the “most obvious” way. (And if you were wondering what made me resurrect and complete a draft on “my session id” a couple of days ago, this posting is the reason). Specifically if you want to select some information for your own session from v$session the query you’re likely to use in any recent version of Oracle will probably be of the form:


select {list for columns} from v$session where sid = to_number(sys_context('userenv','sid'));

Unfortunately that one little statement hides two anomalies – which you can see in the execution plan. Here’s a demonstration cut from an SQL*Plus session running under 19.3.0.0:

connor_mc_d's picture

The definition of proof

One of the pieces of advice that I often see on the ‘net is that undo space is somehow this incredibly precious thing, and as a consequence, one should always keep the amount of uncommitted changes in the database to a small size.

Personally I think that is baloney (Ed-in reality, as an Australian I have a slightly more powerful choice of term, but lets keep things PG-rated </p />
</em></p></div>
    <div class=»

Jonathan Lewis's picture

_cursor_obsolete_threshold

At the recent Trivadis Performance Days in Zurich, Chris Antognini answered a question that had been bugging me for some time. Why would Oracle want to set the default value of _cursor_obsolete_threshold to a value like 8192 in 12.2 ?

In 11.2.0.3 the parameter was introduced with the default value 100; then in 11.2.0.4, continuing into 12.1, the default value increased to 1,024 – what possible reason could anyone have for thinking that 8192 was a good idea ?

The answer is PDBs – specifically the much larger number of PDBs a single CBD can (theoretically) support in 12.2.

In fact a few comments, and the following specific explanation, are available on MoS in Doc ID 2431353.1 “High Version Counts For SQL Statements (>1024) Post Upgrade To 12.2 and Above Causing Database Slow Performance”:

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