Summary2: Directory Name cache

From: Long, Michael (MLONG@karmax.com)
Date: Thu Oct 10 1996 - 08:29:00 CDT


Here's some more info I received after I posted my summary.

        Change these parameters in /etc/system:-
   

                set ncsize = 15000
                set ufs_ninode= 30000
                set bufhwm = 10000

Chnages take effect on next reboot.

Word of warning make a backup of /etc/system first and if it doesn't boot
properly you can boot with the '-a <old-system-file>' and try again.

There are other things you might want to tweak as well. To see what the
dnlc rate is, try the following:

vmstat -s

This will produce a lot of data that you might be interesting

Thanks to the following for your replies:-

Umesh Vaghela
Alex Finkel
Mike (Mehran) Salehi
Scott MacDonald
Garry D. Robbins
Raju Krishnamurthy
Eric Bockemuehl
Richard C. Mills
and Rahul Roy who responded within 30 minutes of my post.

My original question:-

My original question:-

I'm getting the following message on my Sparc station running Solaris
2.5:-

Adrian detected Directory Name cache problem (amber)
Poor DNLC hitrate, increase ncsize
DNLC 78.3%, reference rate 218.48/s
DNLC has 144 entries, try increasing it (and inode cache) to 2288

How do I increase ncsize etc. Sorry for my ignorance but I am new to sun
unix

Regards,

Michael Long.
Karmax Heavy Stamping, Canada

E-mail: mlong@karmax.com



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:11 CDT