Re: # semaphors used by ORACLE

From: Joel Garry <joelga_at_rossinc.com>
Date: Mon, 3 Oct 1994 22:33:58 GMT
Message-ID: <1994Oct3.223358.20133_at_rossinc.com>


In article <35ve19$5p8_at_pandora.sdsu.edu> gtupper_at_monkfish.nosc.mil (Greg Tupper) writes:
>We have an oracle 7 db on an hp 9000/700 and would like to build another. In
>the attempt we get a "Not enough semaphores" error of some sort (I do not have
>the test handy.) I know how to get the number of semaphores on the machine,
>but I do not know how to get the number of semaphors oracle is using.
>
>Anyone have a solution?
>

Well, you can try man ipcs, but that doesn't really answer your question. Although you should check it first, to make sure you haven't accidentally locked up some memory. In other words, ipcs -s should only show you one segment owned by oracle per instance. If you have more, man ipcrm or reboot the machine (note to lurkers - some OS's leave temp files in /tmp or somewhere that show up when you ipcs - rm them in your rc [startup] files if you have that problem).

Probably the easiest thing to do is just increase the semaphore variables that Oracle uses. Don't have a book handy, but I think it's SEMMNS and SEMMNI. You should probably find out what the error is before you put too much work into tuning this. unix virtual memory is very dynamic and to answer your question exactly you would have to write a c program to get the information from the kernel - and IMHO, it's not worth the trouble. Masochists can man semop.

The possibility also exists that you are getting a misleading error message, and it is really a shared memory problem. In that case, you have to deal with SHMMAX and SHMSEG, or make your SGA smaller in the INIT.ORA - or buy more real memory, or upgrade to a bigger box or something. (I see you're _at_ nosc :)

-- 
Joel Garry           joelga_at_amber.rossinc.com            Compuserve 70661,1534
These are my opinions, not necessarily those of Ross Systems, Inc.
%DCL-W-SOFTONEDGEDONTPUSH, Software On Edge - Don't Push.  
panic: ifree: freeing free inodes...
Received on Mon Oct 03 1994 - 23:33:58 CET

Original text of this message