Re: Does IPC help performance?

The 'rdbms ipc message' is the idle event that some of the background processs will wait on. The IPC here is not the same IPC that we see or use on UNIX. The message that is waited on will be placed in the SGA by the foreground and then the foreground will post the background process . When the background process wakes up, it will check its message area in the SGA to see what message is there. Based on that message and the type of process, it will perform some action.

Only the action based background processes will wait on 'rdbms ipc message'. pmon and smon are not action based and wait on their own idle event (pmon timer, smon timer). So you could say that rdbms ipc message is similar to pmon timer and smon timer.

Anjo Kolk.

Jonathan Lewis wrote:

> The RDBMS IPC MESSAGES event is almost always
> down to things like the dbwr 3 second idle write, and
> similar idling activity from the other background processes.
>
> Nothing to worry about. There are various other causes for
> IPC messaging (e.g. segment checkpoints) but you might
> set your mind at rest by demonstrating that the figures for
> this event in v$system_event are very close to the sum of
> the figures for dbwr et. al. in v$session_event.
>
> --
>
> Jonathan Lewis
> Yet another Oracle-related web site: www.jlcomp.demon.co.uk
>
> Sybrand Bakker wrote in message <379753df.983567_at_news.demon.nl>...
> >
> >1. You won't be able to connect as a 'local' user anymore, so you
> >won't be able to stop the database, unless you set remote_os_authent =
> >true and the other usual measures.
> >2. When I was working on Unix sites, my number of waits for RDBMS IPC
> >messages was always high. Research showed it shouldn't be a problem.
> >However, I'm a bit anxious about it too
> >3 background processes are ALWAYS using IPC.
> >
Received on Fri Jul 23 1999 - 03:59:53 CDT