Re: One Extremely Busy Disk

["MDC Information Systems" <mdcis@xxxxxx>]

Disk utilization at this level will cause major response time issues,
queueing theory will prove that anything over 50% on a physical device is
bad (below 30% is ideal, over 70% is death). Look at the hardware error
logs, is there a disk about to fail? are the drives of the same type within
an ASP? what is the disk space utilization on individual drives?
PANIC? yes.

Mike Dunnion   www.mdc.ie

-----Original Message-----
From: prumschlag@phdinc.com <prumschlag@phdinc.com>
To: MIDRANGE-L@midrange.com <MIDRANGE-L@midrange.com>
Date: 23 March 2001 14:55
Subject: One Extremely Busy Disk


>
>
>We have been having an intermittent response time problem lately that
always
>seems to happen around 7:00 AM.  This is shift change time in the factory,
so
>there is plenty of activity on the system (in other words, tough to tie the
>problem to one user or job).  I have looked at the Ops Nav monitors and one
>thing looks out of kilter.  At about this time, one of the disk drives gets
very
>busy for a short period of time, usually 30 to 60 seconds.  This morning
was the
>worst I have seen: a full three minutes at 90% busy.  In the four days that
I
>have watched it, it has always been disk # 21 or 37.  Our journals are in a
>separate ASP, not one of these drives.  I know of no jobs that we have
setup
>that would force a write to the disk for every record update.  That does
not
>mean we don't have any, it just means I may not know about it.  CPU
activity
>looks normal during this time frame.  Disk Utilization is around 80% .
>
>Now that I have rambled on, three more specific questions:
>
>   1.  We don't normally watch disk drive activity, because it just usually
is
>   not a problem so I don't know if this is a normal occurrence or not.  Is
this
>   a red flag?
>
>   2.  How can I tie the activity on the disk back to a specific job
causing the
>   activity?
>
>   3.  Could there be some kind of system activity that is forcing a disk
>   read/update to the same record?  Is it time to get IBM involved?
>
>We are running J D Edwards World on  a V4R3 model 730, with 48 disk drives
(
>6607-070's, 072's, and 074's) with RAID active.
>
>Phil Rumschlag
>
>
>+---
>| This is the Midrange System Mailing List!
>| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
>| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
>| To unsubscribe from this list send email to
MIDRANGE-L-UNSUB@midrange.com.
>| Questions should be directed to the list owner/operator:
david@midrange.com
>+---
>
>

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---