HP3000-L Archives

January 1996, 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:
Curtis Larsen <[log in to unmask]>
Reply To:
Date:
Thu, 25 Jan 1996 14:11:48 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (50 lines)
I just wanted to post a little bit more information about this problem
I'm having with TurboStore II balancing sets using the parallel option.
 
I heard some good advice about the tape drives using hardware
compression, and compressing one set better than another, but that was
with Orbit, and may not be applicable here.  So I went back to the
$stdlist of the full store, chewed on it a little, and came up with
stuff so wild I might just place a Response Center call on it.
 
Set 1 is on LDEV#7, an internal 2000DC (C1504B) tape drive.
It received 64.5 million sectors to store, and used _two_ 90m tapes.
 
Set 2 is on LDEV#8, and external 2000DC (C1521B) tape drive.
It received 17.5 million sectors to store, and used _six_ 90m tapes.
 
So assuming I'm not off my nut -- and that can be a leap of faith at
times -- I would have to say that TurboStore did not exactly balance
things out here, regardless of hardware compression.
 
Now the only numbers I used for these figures were from the SECTORS
column of the store listing, added up via brute force and MPEX's
REPEAT...FORRECS loop -- [Shameless Plug] VESoft: Anybody ever tell you
guys that the world would be a little dimmer without this product? -- so
they should be correctly reflecting the sector size of the files at the
time of the store, and -- as far as I know -- would be the only valid
criteria for balancing load between tape drives.
 
Lastly, the store's progress percentage moves right along at 1% every 5
minutes, slowing only occasionally, until it gets to around 67%
complete, whereupon it changes to a tear-jerking 1% every 20 minutes
until the full store job ends.  This tends to make me think something
I'm storing is making our Operators lose sleep, but when combined with
the above set totals, makes me want to take up mumbling.
 
What I've decided to do for now is to simplify things and just perform a
single-device full store this weekend and examine that for anything
unusual before bothering the HPRC folks.  That should let me know for
certain if anything I'm storing is taking up alot of tape space in
particular.
 
I submit this to both let you know How Things Have Been Progressing, and
to also ask again that if anyone has seen, heard of, or experienced this
problem before, please let me know what you did to resolve it, or if you
see something I'm missing, well hey, just start out the reply with "Boy,
Did You Miss Something..."
 
Again, thanks in advance for any help you can send my way.
---
~ CMPQwk #1.42~ UNREGISTERED EVALUATION COPY

ATOM RSS1 RSS2