I'd
set the shmmin to 1 and the shmmax to <machine memory
size>.
if it
were me, i'd probably have over-allocated something, so, i'd do an ipcs
-a
for my
"couple of databases" and make sure i had no more than two chunks
of
shared
memory allocated ( ignoring root stuff, of course). Not that
breaking
the
SGA into a few pieces is a crime against humanity, but....no positive
need
to do
it, either....
the
unit shmmax is measured in is OS page sizes as i recall, but...just RTFM
on
/etc/system and it'll make it clearer than i could.
hth
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, February 27, 2001 5:26 PM
To: Multiple recipients of list ORACLE-L
Subject: /etc/system parameters on Solaris 8Hey everyone,Recently I have setup a couple 8.1.6 databases on Solaris 8. The databases appear to function normally, but when I shutdown either of them I receive the following message in a trace file:skgm warning: EINVAL creating segment of size 0000000004100000
fix shm parameters in /etc/system or equivalent(the message is repeated several times except for the segment size being different)I checked the /etc/system parameters and they are the same as some of my other boxes (Solaris 6 machines running 4-6 databases each) and they seem to work fine. Memory shouldn't be a problem on this machine since it has 4 Gb of RAM with only 2 databases up and running. Running 'top' shows that plenty of memory is available also. Here are the /etc/system parameters for Oracle....set semsys:seminfo_semmni=750
set semsys:seminfo_semmsl=600
set semsys:seminfo_semmns=1200
set shmsys:shminfo_shmmax=67108864
set shmsys:shminfo_shmmni=1200
set shmsys:shminfo_shmseg=180
set shmsys:shminfo_shmmin=30Does Solaris 8 require some of these parameters to be higher? I tried looking on MetaLink, but....well, you know that story. Ask Ross Mohan for the details on Metalink....Thanks,AlanAlan AschenbrennerOracle DBAIHS Group(303)-858-6394
