ADSM-L

Re: Re[2]: Anyone else having server database contention problem

1998-06-09 09:58:45
Subject: Re: Re[2]: Anyone else having server database contention problem
From: "Purdon, James" <james_purdon AT MERCK DOT COM>
Date: Tue, 9 Jun 1998 09:58:45 -0400
Hi again,
  I'm still interested in your configuration. In particular, the size of
your ADSM database, how its laid out on disk, and the total number of
clients you are backing  up with it.  Also, how long have you been using
ADSM to backup that particular system (and others) and the amount of change
on  your systems.

  I think that my suggestion that database search times might be dependent
on client hardware type is incorrect.  It is more likely that the
differences I observed are due to the size of the ADSM DB at each instance
and the client's backup history (the AIX restore occured shortly after ADSM
was installed, the Cray restores after about 9 months of operation and the
SGI restore after two years).

Normally, I try to keep the paranioa under control.
Jim

> ----------
> From:
> Carl.Makin AT ipaustralia.gov DOT au[SMTP:Carl.Makin AT ipaustralia.gov DOT 
> au]
> Sent:         Monday, June 08, 1998 8:42 PM
> To:   ADSM-L AT VM.MARIST DOT EDU; Purdon, James
> Subject:      Re[2]: Anyone else having server database contention problem
>
>
>
>
>
> > Author:  "Purdon James" <james_purdon AT MERCK DOT COM>
>
> > Guess again.  It would probably take two weeks just to get a listing of
> 9.5
> > million files in a single filespace.
>
> 53 filespaces, not 1.  One attempt at restoring a 15Gb filesystem took 8
> hours.  It has 390,000 images in it.
>
> >> From:         Carl Makin[SMTP:Carl.Makin AT IPAUSTRALIA.GOV DOT AU]
> >>> Author: Bill Quintrell <Bill_Quintrell AT PROVIDENTCOMPANIES DOT COM>
>
> >>> Not to start a contest (our record is 52 hours for one incremental
>
> >> I killed one of our backups after 4 days, it had done around 60% of
> > 400Gb containing 9.5 Million 40Kb(average) TIFF images in 53 filesystems
>
>
>
> Carl.
>
>
<Prev in Thread] Current Thread [Next in Thread>