From: Louis Avrami (avramil@concentric.net)
Date: Tue Nov 18 2003 - 02:12:20 EST
Hello all,
I'm working on a Solaris 2.8 system which has just various security
hardening techniques applied to it, including using the coreadm
command to name core dumps uniquely and redirect them to a central
location, /var/core.
On our systems the /var partition isn't too large, about 2 gig
total, perhaps 780 meg of free space. This server hosts an Oracle8i
database. Unfortunately, there is a RDBMS bug which causes a
core dump when the bug is tripped over. The core dump file is
quite large, approximately 200 meg. After a few core dumps,
/var fills up. The nature of the environment is that this probably
wouldn't be noticed for awhile, or until something failed.
The UNIX admin is of the opinion that /var running out of free
space isn't a big deal. I'm worried about things like vi not
working, syslogd, crond, utmp, wtmp, and other things not being
able to function properly and the effects over a period of time.
I've suggested either creating a separate non-system partition,
/core, or making /var/core its own partition, just to avoid any
potential trouble.
Am I being paranoid? Can anyone suggest a "best practice" for
redirecting core dumps with coreadm on a large server?
Thanks,
Lou Avrami
_______________________________________________
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 - 23:27:30 EDT