HP3000-L Archives

March 2001, Week 2

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:
John Clogg <[log in to unmask]>
Reply To:
John Clogg <[log in to unmask]>
Date:
Fri, 9 Mar 2001 10:08:38 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (73 lines)
You have a couple of pages of memory that have been deallocated due to
errors.  You should contact your hardware maintenance provider about it.
You could almost certainly benefit from more memory.  Whether that is the
primary cause of your performance problems is hard to tell from the
information you provided.  You said you have page fault rates above 50/sec.
How far above 50?  Another indicator of memory pressure is the number of
I/O's per second on behalf of the memory manager.  Numbers significantly
above 20 are considered bad.  The "mem" wait state shown for a process is
actually "memory prefetch", which is a disc I/O wait.  The 12H can certainly
be a performance killer, depending on how it's configured.  Depending on how
your data is organized and accessed, more memory can do a lot toward
alleviating disc bottlenecks by achieving better cache hit rates.

-----Original Message-----
From: DeLisa DeVee [mailto:[log in to unmask]]
Sent: Friday, March 09, 2001 10:07 AM
To: [log in to unmask]
Subject: Memory HP3000 979-300 / System Performance


We are having some system performance issues.  We are questioning if our
system is using all of its memory, or if we have some bad memory.  Below
is a print out of the memrpt in sysdiag/logtool.  This report shows
error count on ib - 2 and 1ab - 1.  Does anyone know if this is normal?

We are also trying to determine if we have enough memory.  According to
HP Response Center we do, but we continue to have page fault rates above
the normal recommendations (above 50.00/sec) and have users and
processes waiting on memory.  We currently have 896 mb of physical
memory.  We are on operating system 5.5.7 and have 250 users on this
box.

We recently added a third processor and our system performance declined
tremendously. We also have a 12H AutoRaid on this system.  We are trying
to determine why our system performance is so poor (locking, memory
etc.)?  Any suggestions would be greatly appreciated.

                 MEMORY ERROR LOGGING REPORT

SYSTEM START TIME             --  TUE, MAR  6, 2001,  6:58 PM
LAST ERROR CHECK TIME         --  FRI, MAR  9, 2001,  9:27 AM
LOGGING TIME INTERVAL         --  3600 seconds
49 OUT OF 50 ENTRIES IN THE PAGE DEALLOCATION TABLE ARE FREE

MEMORY EXTENDER  --  EX0
 ===========================================================================

SLOT:               1b                  FIRST DETECTED:    03/06/01
7:58 PM
ERROR TYPE:         Single/Soft Error   LAST DETECTED :    03/06/01
10:58 PM
PAGE STATUS:        Deallocated         CHIP ID:           u14
BIT NUM./BANK NUM:  97/0                ERROR COUNT:       2
LOGGED BY:          MEMLOGP             ERROR ADDR (HEX):  117E9810
 ===========================================================================

SLOT:               1ab                 FIRST DETECTED:    03/06/01
6:58 PM
ERROR TYPE:         Single/Unknown Err  LAST DETECTED :    03/06/01
6:58 PM
PAGE STATUS:        Deallocated         CHIP ID:           Not Available

BIT NUM./BANK NUM:  Not Available       ERROR COUNT:       1
LOGGED BY:          PDT                 ERROR ADDR (HEX):  117E9000
 ===========================================================================



DeLisa DeVee
Operations Administrator
Musician's Friend
541-774-5473

ATOM RSS1 RSS2