RE: linux kernel parameters
Date: Sat, 13 Feb 2021 12:13:24 -0800
Message-ID: <2e4201d70244$af34e530$0d9eaf90$_at_comcast.net>
It is not about being right, it is about being technically accurate. You make statements with no facts.
You say that the values imply you can only have 11 semaphore sets but the definitions of the parameters and the values provided say otherwise
3000 32000 3000 128
kernel.sem = SEMMSL SEMMNS SEMOPM SEMMNI
3000: semmsl - Maximum number of semaphores per set 32000: semmns - Maximum number of semaphores on the entire system 3000: semopm - Maximum operations for semop system call 128: semmni - Maximum number of semaphore arrays on the whole system
SEMMNI shows the whole system is set to 128 semaphore sets/arrays available not 11 as you have stated. Then you use the 11 value as justification to state the other parameter settings are wrong.
Also examining Oracle built systems with Oracle built parameters not changed by the customer (meaning Oracle recommended settings even if not in the docs) Oracle uses values that are different than you recommendation:
Exalytics X5: kernel.sem = 2048 64000 256 128 Exadata X6-2: kernel.sem = 1024 60000 1024 256 Exadata X8-8: kernel.sem = 2048 400000 256 256 Exadata X8-2: kernel.sem = 1024 60000 1024 256 Exadata X5-8: kernel.sem = 2048 262144 256 256 Exadata X5-2: kernel.sem = 1024 60000 1024 256
In every case the Oracle set values do not match your statement: " The problem is that SEMMNS has to be larger or equal than SEMMSL*SEMMNI. Ideally, SEMMNS=SEMMSL*SEMMNI and SEMMSL=SEMOPM."
As a matter of fact the Oracle provided settings more follow this formula SEMMNS < SEMMSL*SEMMNI and SEMMSL >= SEMOPM
It all depends on what work is being performed on the system and what the requirements of the applications are running on the system.
-----Original Message-----
From: Gogala, Mladen <gogala.mladen_at_gmail.com>
Sent: Saturday, February 13, 2021 11:16 AM
To: dimensional.dba_at_comcast.net; oracle-l_at_freelists.org
Subject: Re: linux kernel parameters
Actually, it does. It confirms the syntax of the semaphore parameters. However, you seem to have the insatiable need to be right, so I'll not disturb you any longer. You don't seem to cope well with the Covid isolation. Enjoy the rest of it.
On 2/13/2021 1:50 PM, dimensional.dba_at_comcast.net wrote:
> The information I presented does not agree with you either.
-- Mladen Gogala Oracle DBA Tel: (347) 321-1217 Blog: https://dbwhisperer.wordpress.com -- http://www.freelists.org/webpage/oracle-lReceived on Sat Feb 13 2021 - 21:13:24 CET