who has done this very, very risky manipulation yet ?

From: rob.de.langhe@belgacom.be
Date: Thu Jun 30 2005 - 12:04:54 EDT


Hi,

our Oracle admins want now and then add a new piece of database server
software, but hit the limits on Shared Memory parameters, like shmmax.

I understand that this is a value to block processes from requesting
higher values, but should for the rest have no direct relation with
kernel table sizes, etc.

To avoid that we need to go always through a server reboot after
adapting the value in /etc/system, I understand from SUN that with "adb"
you get modify the value of such parameter in the kernel while the
system is running, but at the same time they categorise this
manipulation as very very risky business.

My idea is: either it is supposed to work, and thus NOT risky, or it is
NOT supposed to work hence risky if you do it anyhow.

Who has done such manipulation already on a live server, what's your
experience (simply stated: paniced system ?), and does anyone know about
another way to adapt kernel tunables without having to go through all
those reboots. Hey, we want to have different sysadmin methods than the
M$ competitor, no ?

TIA

Rob

**** DISCLAIMER ****
http://www.belgacom.be/maildisclaimer
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:31:00 EDT