Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site ucbvax.BERKELEY.EDU
Path: utzoo!decvax!ucbvax!HARVARD.HARVARD.EDU!sasaki
From: sasaki@HARVARD.HARVARD.EDU (Marty Sasaki)
Newsgroups: mod.computers.vax
Subject: Directory cache hit rate, the solution
Message-ID: <8610230257.AA23595@ucbvax.Berkeley.EDU>
Date: Wed, 22-Oct-86 22:45:58 EDT
Article-I.D.: ucbvax.8610230257.AA23595
Posted: Wed Oct 22 22:45:58 1986
Date-Received: Fri, 24-Oct-86 04:48:38 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 18
Approved: info-vax@sri-kl.arpa

We were having problems with a very low directory cache hit rate even
though we had a large cache. I didn't mention it in my note, but we
have another system in the cluster which had identical hardware, a
smaller directory cache, but a much higher hit rate. Everyone was
puzzled.

The solution was simple, allocate more pool (I can hear all of the RSX
guys saying, "of course, allocate more pool..."). Turns out that when
the ACP goes to allocate the caches and finds that it can't get enough
memory, it reverts to a small default and tries to allocate that.

Yes, we did use AUTOGEN to set things up. Don't worry DEC, the SPR is
in the mail...
----------------
  Marty Sasaki				uucp:   harvard!sasaki
  Strategic Information			arpa:	sasaki@harvard.harvard.edu
  80 Blanchard Road			bitnet: sasaki@harvunxh
  Burlington, MA 01803			phone:	617-273-5500