Re: local write wait event

  • From: "Jonathan Lewis" <jonathan@xxxxxxxxxxxxxxxxxx>
  • To: <chupit@xxxxxxxxx>, <oracle-l@xxxxxxxxxxxxx>
  • Date: Tue, 12 Oct 2004 21:05:40 +0100

The only time I have seen local write wait
and tracked it back to source, it has been
the query co-ordinator cleaning up after
parallel execution slaves have been involved
in parallel create table / index.  Each slave
thinks it owns a segment, so creates a segment
header as the first block of the data it generates.
The QC wipes all but one of them and puts them
into the segment free list.  

There may be other reasons that I've not yet
come across.


Jonathan Lewis
The Co-operative Oracle Users' FAQ
Optimising Oracle Seminar - schedule updated Sept 19th

----- Original Message ----- 
From: "Edgar Chupit" <chupit@xxxxxxxxx>
To: <oracle-l@xxxxxxxxxxxxx>
Sent: Monday, October 11, 2004 7:58 AM
Subject: local write wait event

Dear List,

Tracing one of the user sessions I get a lot of waits on "local write
wait". We are using 10.0.3, I've searched metalink and in the book
about OWI events, but couldn't find any information about this event,
while searching in the archives I've found post by K Gopalakrishnan,
that this event should never be seen in normal databases. Maybe there
is a more detailed description of this wait event?



Other related posts: