HP3000-L Archives

March 1996, Week 1

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:
Mark Ranft <[log in to unmask]>
Reply To:
Mark Ranft <[log in to unmask]>
Date:
Mon, 4 Mar 1996 15:13:51 GMT
Content-Type:
text/plain
Parts/Attachments:
text/plain (60 lines)
In response to numerous user requests HP changed the disc allocation mechanism
in MPE/iX Release
5.0.  The intent was to keep from fragmenting the disc space on LDev 1.
 
Last week I discovered that new algorithm does not work on systems that have
less than four(4)
members(discs) on the MPEXL_SYSTEM_VOLUME_SET.  In this case, LDev 1 is almost
never used for
file allocation.  This includes virtual memory, permanent file space as well as
spoolfiles.
In response to numerous user requests HP changed the disc allocation mechanism
in MPE/iX Release
5.0.  The intent was to keep from fragmenting the disc space on LDev 1.
 
Last week I discovered that the new algorithm does not work properly on systems
that have less
than four(4) discs on the MPEXL_SYSTEM_VOLUME_SET.  In this instance, LDev 1 is
almost never
used for disc space allocation.  This includes virtual memory,
temporary/permanent file space as
well as spool files.
 
On one of our systems, Oracle locked up and crashed because LDev 2 ran out of
space.  (I had
over 3 million sectors available on LDev 1 at the time.)
 
This opportunity is reported in SR# 5003 268938.
 
A beta patch is being produced to correct the problem.  (It may or may not make
it into Release
5.5.)  The beta will be some weeks to come.  An emergency 'work around' (a
DEBUG 'NOP' of the
call to the algorithm) was offered to me.  Instead of installing the patch, I
decided to move
some files to LDev 1 to free up space on LDev 2.
 
To determine if the new algorithm is affecting you, use "DISCFREE C" and look
at LDev 1
allocation (virtual memory allocation is a sure indicator.)  Mine looks like
this...
 
           |    Configured     |      In Use       |     Available     |
-----------+-------------------+-------------------+-------------------+
 
LDEV :     1 -- (MPEXL_SYSTEM_VOLUME_SET:MEMBER1)
 Device    |    5294144        |    3229856 ( 61%) |    2064288 ( 39%) |
 Permanent |    3970608 ( 75%) |    3122096 ( 59%) |     848512 ( 16%) |
 Transient |    3970608 ( 75%) |     107760 (  2%) |    2064288 ( 39%) |
 
LDEV :     2 -- (MPEXL_SYSTEM_VOLUME_SET:MEMBER2)
 Device    |    5294144        |    3387168 ( 64%) |    1906976 ( 36%) |
 Permanent |    5294144 (100%) |    2164928 ( 41%) |    1906976 ( 36%) |
 Transient |    5294144 (100%) |    1222240 ( 23%) |    1906976 ( 36%) |
 
Note how only 2% of LDev 1 is being used for Transient (Virtual Memory).
 
Mark Ranft
General Mills Inc.
[log in to unmask]

ATOM RSS1 RSS2