I have 5 instances running on an oracle solaris 2.8 server. Each with the following setup

inst-a = 200 processes
inst-b = 200 processes
inst-c = 100 processes
inst-d = 50 processes
inst-e = 50 processes

My system set up is

set semsys:seminfo_semmni=100
set semsys:seminfo_semmsl=510
set semsys:seminfo_semmns=1030
set semsys:seminfo_semopm=100
set semsys:seminfo_semvmx=32767

When I do check the semaphores on the system, I notice

s 196608 0xd8ded9a8 --ra-r----- 204
s 196609 0xfcef8940 --ra-r----- 71
s 1376258 0x109ccca8 --ra-r----- 204
s 196611 0x834c0788 --ra-r----- 54
s 196612 0xd87dff2c --ra-r----- 104

Now comes the worst part. When I tried to increase the processes of inst-a to 500 and restart the instance, I'm getting ORA-21746. This was the setup some-time back. :D Why it is so, even if we try to do the math over here

say suppose that my new instance consumes 510 NSEMS, then the total would only come to 943. Where my SEMMNS is 1030. So the new setup would cleanly fall within the limits. I'm now confused of why oracle is poping out this ORA-217146.

I know that just bumping the semmns, would fix this problem, but still couldn't get why? :confused: :confused:

Any ideas, or thoughts????