Re: ORA-07279 using ORACLE 8.0.5 and HPUX 10.20, semaphore set error

Steve:

Bear with me, I have a cold and I don't have my notes. Search this group for my article on Oracle on NT vs. Solaris.

There are three semaphore parameters that are interrelated. The first is the number of semaphore sets, the second is the number of semaphores/set, and the third, and this is the kicker, is the total number of semaphores in the system, which must be the multiple of the first two. If your total semaphores is less than the number of semaphores in each set HP UX won't be able to even create the first semaphore set.

Take care, e-mail me if my reply is too confused to make sense out of.

Nigel

In article <369B96A2.2AB_at_worldnet.att.net>,   Steve Maestas <smaestas_at_worldnet.att.net> wrote:
> If anybody is running Oracle and HPUX and could give me a rough idea on
> what to set the Semaphore parameters to, I would greatly appreciate it.
> I am running 4 Oracle instances v8.0.5 on a HPUX v10.20 box, with 1.5
> Gig RAM.
> I initially set the kernal parameter, semaphore specific, to the
> following:
>
> SHMMAX 1073741824
> SHMMIN 100
> SHMMNI 600
> SHMSEG 60
> SEMMNS 800
> SEMMNI 280
>
> This worked fine, but when I modified the ora files to enable "large"
> databases settings, I was able to stop/start the first instance. When
> I attempted to stop/start the second instance I recieved the following
> error message:
>
> ORA-07279: spcre: semget error, unable to get first semaphore set.
> HP-UX Error: 28: No Space left on device
> Additional information: 200
> Additional information: 2
>
> I bumped up the semaphore parameters and rebooted. The instances
> started up fine, but when I changed a third instance for "large" db
> settings, I recieved the same message. To test I stoped all Oracle
> instances and attempted to start the "changed" instance, same error
> message. I do not understand the whole semaphore concept, but it
> seems to me if I stoped the other instances the semaphores would be
> free, and I would be able to start the "changed" instance. If I
> now attempt to start the other instances I get this message. I am sure
> if I reboot, it will fix the problem and the instances will be running.
> Is there a bug or fix for either Oracle or HPUX that fixes this, or is
> this just how it works and my own ignorance?
>
> Current Settings:
>
> SHMMAX 1073741824
> SHMMIN 100
> SHMMNI 600
> SHMSEG 60
> SEMMNS 1000
> SEMMNI 400
> SEMMNU = SEMMSL = 1000
>
> Thanks
>
> Steve
>

-----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own Received on Tue Jan 12 1999 - 16:57:13 CST