HP3000-L Archives

January 2005, 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:
Christian Lheureux <[log in to unmask]>
Reply To:
Christian Lheureux <[log in to unmask]>
Date:
Fri, 21 Jan 2005 09:53:22 +0100
Content-Type:
text/plain
Parts/Attachments:
text/plain (67 lines)
Hi Ernie,

The seemingly odd behavior of LDEV 1 runs back to an old enhancement brought into 5.5 (IIRC). It consists of, among all MPEXL_SYSTEM_VOLUME_SETS, of always using LDEV 1 in last order when trying to allocate PERMANENT space. This has been done to make sure that ANY MPE system always has sufficient TRANSIENT space to boot. As a result, LDEV 1 of course has less PERMANENT space than the other system disks so it can harbour more TRANSIENT space, i.e. virtual memory (as correctly reminded by Denys).

The 95% PERMANENT and TRANSIENT space limit you quote are not a problem per se. MPE would never let you allocate a 100% PERMANENT limit. It will always make sure ANY disk used as LDEV 1 has at least 360,000 sectors, which are needed, IIRC, during boot time as permanent space. In theory, you could even stretch your LDEV 1 to, I think, something like 97 %, which would still be within MPE bounds or, more precisely, Volume Management. Any attempt to configure such a high PERMANENT space limit to something that does not provide at least the sacrosanct 360,000 TRANSIENT sectors needed during boot will result in an error 6 in VOLUTIL. This is NOT the 5.5 enhancement. IIRC, this feature of transient space conservation during boot is as old as MPE/XL itself, i.e. mid-80s. It was very useful at a time when the standard disk was the 7935, which was exactly 1,559,704 sectors in space. Hence the "traditional" 75% limit (a default config value, I think), which is close to 75 % of the size of the old 7935.

I wrote an article about that at the time I beta-tested 5.5. I'm now quoting from a 10-year old memory, with a few burned-out neurons since then. Please pardon the inaccuracies, if any. Gee, if I could only get my hands on my article !!!!!

Christian Lheureux
Directeur BU Infrastructure / Manager of IT Infrastructure BU
________________________________


  29, rue de Rosières
  91240 SAINT-MICHEL-SUR-ORGE
  FRANCE

  Téléphone :
+33 1 69 72 12 83
  Fax :
+33 1 60 15 30 62
  Email :
[log in to unmask]
  Web : 
www.btwcomputing.com
________________________________

En mai 2004,APPIC RH devient BTW Computing, une activité du Groupe BTW
In May 2004, APPIC RH becomes BTW Computing, an activity of the BTW Group 
> -----Message d'origine-----
> De : HP-3000 Systems Discussion [mailto:[log in to unmask]] De la part
> de Newton, Ernie
> Envoyé : jeudi 20 janvier 2005 20:35
> À : [log in to unmask]
> Objet : [HP3000-L] Transient space on drive 1
> 
> Greetings,
> 
> Along with security issues, Vladimir noted that there was
> too much transient space used on our drive 1. We utilize a
> mod 20 disk array, with drive one consisting of 2 4-gig
> drives.  Here's what my DISCFREE shows...
> 
>            |    Configured     |      In Use       |     Available     |
> -----------+-------------------+-------------------+-------------------+
> 
> LDEV :     1 -- (MPEXL_SYSTEM_VOLUME_SET:MEMBER1)
>  Device    |   16411488        |   11282640 ( 69%) |    5128848 ( 31%) |
>  Permanent |   15590912 ( 95%) |    9946032 ( 61%) |    5128848 ( 31%) |
>  Transient |   15590912 ( 95%) |    1336608 (  8%) |    5128848 ( 31%) |
> 
> 
> My other drives show in-use transient space at zero or one percent.
> I've
> never really understood exactly what this means, and we have NO problems
> with performance.  So...is this a bad thing?  Should I be chastised?
> 
> Thanks,
> 
> Ernie
> 
> * To join/leave the list, search archives, change list settings, *
> * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

ATOM RSS1 RSS2