Since this comparing is one of the keys to Retrospect's ease of use
(something I wouldn't want to give up), I hope you are considering
ways to make it faster. I'm only going to be wanting to back up more
clients with more files (most likely a lot of small ones rather than a
few big ones) in the future. Let's see, with my 15 clients (some with
>1 volume), and 6 backups a week (one "full" and 5 "normal"), I get
over 100 sessions before I move to a new set. Even on the first
"normal," the times seem rather long for the cataloging processes
(almost double the first night, but only slightly longer subsequent
nights). Here are the totals from last week's backups (I do a New
Media backup every Friday night).
Fri Total duration: 1 05:03:16 (05:13:13 idle/loading/preparing)
Sun Total duration: 10:24:51 (10:02:29 idle/loading/preparing)
Mon Total duration: 10:59:40 (10:23:30 idle/loading/preparing)
Tue Total duration: 11:55:39 (10:50:05 idle/loading/preparing)
Wed Total duration: 11:50:30 (10:47:08 idle/loading/preparing)
Thu Total duration: 11:37:20 (10:40:56 idle/loading/preparing)
As you mentioned, I don't think upgrading/adding hardware is a good
option. I'm already running my backups on a 700 MHz machine with
256MB of RAM (and on a dual 600 before that it wasn't any different).
Splitting the clients up into groups defeats one my goals of having
backups be EASY to maintain.
How about when the backup process starts at the scheduled time, you
let the clients get the catalog information from the server, and set
them on their way to do the processing/comparing/etc. and then tell
the server "Hey, I'm ready to send some data." Then the server could
concentrate (mostly) on writing data to tape (or other media) while
the CPU intensive stuff happens at the workstations.
...Doug
-----Original Message-----
From: Eric Ullman
Sent: Tuesday, December 05, 2000 10:05 AM
<snip>
This matching process can take some time. It's dependent upon the
total
number of files in those sessions and the speed of the backup
computer. At a
certain point, the scanning and matching process will take longer than
actually copying the files to the backup set. Scanning and matching
performance is dependent on the total number of files; copy
performance is
dependent on the amount of data being moved.
<snip>
--
----------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Archives: <http://list.working-dogs.com/lists/retro-talk/>
For urgent issues, please contact Dantz technical support directly at
[EMAIL PROTECTED] or 925.253.3050.
