Re: constant monitoring and enqueues?

I think you are likely to kill your system if you monitor every 5 seconds - unless your system is massively overpowered and you avoid some of the larger X$ objects.

You might like to try the trick which I think Precise use - start with x$ksmfsv, which is the fixed memory map for the rest of the SGA, then use offsets into the SGA dictated by the contents of that to get to the x$ structures underpinning the v$ views.

You'll have to write it in C or some other 3GL, but it shouldn't be impossible to get some useful information, and you'll bypass a lot of complex Oracle code doing it, so your access rate can be very high.

Can you give some clues about your enqueue problem - it won't be possible to 'grab them at enqueue time' because that would require a method of intercepting the Oracle code path, but if you check v$session_wait for 'enqueue' waits, that gives you some detail about the enqueue you are waiting for, and x$ksqst gives you some information about the general cost of enqueue waits.

--
Jonathan Lewis
Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk

Practical Oracle 8i:  Building Efficient Databases

Publishers:  Addison-Wesley
See a first review at:
http://www.ixora.com.au/resources/index.htm#practical_8i
More reviews at: http://www.jlcomp.demon.co.uk/book_rev.html



andrew_webby at hotmail wrote in message
<979035313.22136.0.nnrp-12.c30bdde2_at_news.demon.co.uk>...


>Thanks Ethan. I'll take a look when the link comes back up.

>

>I still really need to know about constant monitoring however - like every

 5


>seconds at most. Like I say, this is really just for an exercise more than

>anything else.

>

>Also, I'm still looking for info on grabbing the information from

>problematic enqueues as they occur if anyone else can help?

>

>Andrew