HP3000-L Archives

March 2004, Week 4

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:
Date:
Fri, 26 Mar 2004 11:56:55 +0100
Content-Type:
text/plain
Parts/Attachments:
text/plain (82 lines)
Craig, you are right about XM. The system and data logs have been split in
the good ole' days of 4.0, into a 16-MB system and a 64-MB data log. At the
same time, the recovery processes have been made parallel, thus saving a
helluva time when rebooting a many-spindle config and taking advantage of MP
servers.

Christian

> -----Message d'origine-----
> De : HP-3000 Systems Discussion [mailto:[log in to unmask]]De la
> part de Craig Lalley
> Envoye : vendredi 26 mars 2004 11:45
> A : [log in to unmask]
> Objet : [HP3000-L] internals, and patches
>
>
> Everyone,
>
> There used to be a patch on MPEiX/6.5 that would reduce the
> overhead of the
> transaction manager as it scanned through memory.  I believe
> the patch reduce
> the overhead of semaphore locking during the scan.  I thought
> the patch was
> MPEXMX34A?
>
> I can't seem to find it in the patch database.  Also, I was
> wondering if this
> patch was available in MPEiX/7.0.
>
> Any help would be appreciated.
>
> -Craig
>
> Also, a while back Mr. Denys attempted to correct me and my
> beliefs regarding
> the XM.  He was wrong.
>
> Mr. Denys there are actually two independant XM logfiles per
> volume set, and I
> am not talking about two log halves.  There is a system XM
> logfile witch is
> considerable smaller than the "user" log file which defaults
> to 64MB (32MB
> halves).
>
> The system XM logfiles keeps track of all the system stuff
> like directory
> structure etc... The user XM logfiles keep track of all the
> user stuff, mostly
> IMAGE transactions.
>
> I meant to correct Mr. Denys before today, but I just don't
> have the time,
> between reading all the OT posts.
>
> The XM log file is a write cache that posts when the
> checkpoint is reached.
> There are certainly other ways to instigate a checkpoint, no
> question.  The
> problem on heavy data entry systems it is possible for the
> logfile to fill up
> before the previous XM checkpoint has a chance to finish/post
> (this is a bad
> thing) and it is called a collision.  There are ways to work
> around this, and a
> really fast disc subsystem can help.
>
>
>
>
> __________________________________
> Do you Yahoo!?
> Yahoo! Finance Tax Center - File online. File on time.
> http://taxes.yahoo.com/filing.html
>
> * 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