From time to time I see people asking about the impact of setting the processes parameter to allow for a much larger number of Oracle processes than is really needed to support their application. I must have addressed various aspects of this question seven or eight times on the Internet, but haven’t yet got around to writing anything about it on my blog.
I’ve just addressed the topic yet again on the OTN database forum, so I thought I’d take advantage of the notes I’ve made there by referencing them from here. You might want to read the whole thread, but I’ve supplied links to two posts – one about the effects of the parameter on what goes on in the SGA and one on a follow-up question on how the parameter can affect O/S memory usage if the number of processes running actually jumps to take up the excess that you’ve allowed.
There are probably a few more details worth a mention – so this posting may get updated some time in the future.
Thank you for this information, it’s very helpful.
In one of those posts, you point out that storms of creation and destruction of processes can be very painful. We’ve seen issues in our application with the creation of parallel query processes causing us to hit the limit enforced by the processes parameter, Does the creation of parallel query worker processes cause the same kind of process storms that you say would be good to avoid? Would it be wise to have a small range between PARALLEL_MIN_SERVERS and PARALLEL_MAX_SERVERS to avoid this?
Thanks!
Comment by Ben — September 23, 2009 @ 11:21 am UTC Sep 23,2009 |
[...] http://jonathanlewis.wordpress.com/2009/09/22/processes/ [...]
Pingback by High Procesess & Sessions « Anand's Blog — September 24, 2009 @ 11:29 am UTC Sep 24,2009 |