Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Information on contents of v$system_event view

Re: Information on contents of v$system_event view

From: <ozziboy_at_pacbell.net>
Date: Thu, 23 Apr 1998 21:25:27 -0700
Message-ID: <35401437.7D4CA6F2@pacbell.net>


Oracle got the names wrong,

The way I look at tuning oracle is that "everything is always waiting". therefore always has a wait event. A session is either waiting for something to do or waiting for a response from another event type such as "db file sequential read" in your case. This event means your process is waiting for index reads to be performed. If you were doing full table scans the event would be "db file scattered read". There are not a lot of events that normally need to be looked at. Some other ones are enqueue, free buffer wait, buffer busy wait.

I thought Oracle had documented these in one of the manuals. So it may be a case of RTFM. The solution to your db file sequential read problem depends on what you are attempting to do. You may need to alter the SQL statement, revise the index usage, not use indexes at all, increase the size of the buffers, improve the I/O subsystem on your computer (striping, faster disk, controllers, etc) or a combination of all of these.

Regards,
Andrew.
Quest Software.

John Vernes wrote:

> I have a question regarding the rows in the v$system_event view.
> Does any one have documentation on what to do with the info from this view.
> I know where to find the description of the view and it's columns, but I
> want
> to know what the events mean.
>
> I especially have a question about the 'db file sequential read' row
>
> The total time_waited for this event is extremely high when doing a couple
> of selects. Example: I run two select statements which run over a total time
>
> of an 1/2 hour. After reading the parameter, it states that it has been
> waiting for
> 1200 seconds (20 minutes) and an average wait of 0.02.
>
> The selects are done almost only by reading indexes and we almost certainly
> have an I/O bottleneck.
>
Received on Thu Apr 23 1998 - 23:25:27 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US