HP3000-L Archives

May 1995, Week 3

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"Rudderow, Evan" <[log in to unmask]>
Reply To:
Rudderow, Evan
Date:
Mon, 15 May 1995 09:04:00 EDT
Content-Type:
text/plain
Parts/Attachments:
text/plain (38 lines)
Guy Smith <[log in to unmask]> wrote:
 
 
<snip>
 
>1) Has anyone noticed an marked increase in disc I/Os by
>   Memory Manager under 5.0?
 
<snip>
 
>Heres the issue:  Weve taken a number of system to 5.0.
>One system has had (a) no hardware changes, (b) no
>application software changes, (c) no change in transaction
>volume, (d) no changes in configuration or file placement.
>This systems also (e) logs data to a small set of files
>and thus there is an amazingly high data page residency
>rate.
>
>Oh.  BTW, disc I/Os for memory manager doubled after the
>update.  For this machine, where disc I/Os are typically
>low, its not really a problem.  But our 995, which is I/O
>bound to begin with, will likely not react well to the
>increased load.  I interested in any insights anyone might
>have as to the changes to memory management that would
>double the I/O rate.  Since this machine relies heavily on
>message files (that went native in 5.0), I am tempted to
>target them as suspects, though I would expect a decrease
>in I/Os, not the reverse.
 
This is really unlikely, but...
 
I recall seeing something about a patch to increase performance on systems
with small memory configurations -- perhaps you should try installing that
patch.  On the other hand, if you *have* installed that patch, then perhaps
you should try *de-installing it*...
 
 -- Evan

ATOM RSS1 RSS2