Networker

[Networker] Recovering from multiple savesets in parrallel

2005-06-09 14:16:23
Subject: [Networker] Recovering from multiple savesets in parrallel
From: Craig Ruefenacht <Craig.Ruefenacht AT US.USANA DOT COM>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Thu, 9 Jun 2005 12:11:09 -0600
Hi,

I may not be remembering correctly, but I think I remember being able to
perform a restore from multiple savesets at once.

For example, if I had a host which had 6 filesystems (each a saveset),
say, /u01 - /u06.  When they are backed up, the savesets are multiplexed
to a tape drive.  In my case, upto 4 savesets from an individual host
can be multiplexed at once.

When recovering files from those 6 savesets using the recover command-
line interface:

        recover
        recover: Current working directory is /
        recover> add /u01
        ...
        recover> add /u02
        ...
        recover> add /u03
        ...
        recover> add /u04
        ...
        recover> add /u05
        ...
        recover> add /u06
        ...
        recover> recover
        

It used to be that Networker would recover files from multiple savesets
at once (for instance, it would recover files from /u01, /u02, /u03,
and /u04 at the same time).  Currently, however, Networker recovers them
serially - it will recover files from one saveset (say, /u01), and, when
it completes that, it will start recovering files from the second
saveset (say /u02), and so on.

I am not doing a saveset recover - I am browsing the savesets and
selecing a subset of files from each saveset (filesystem).  Each
filesystem was backed up at the same time and muxed together, 4 at a
time, to tape.

This is when recovering data from a full backup.  Our Networker server
is currently v7.1.3 on HPUX 11i, and clients that I've seen this
behavior on are all Networker v7.1.3 on Linux Redhat and Solaris 5.8.

We recently upgraded Networker from v6.1.3 to v7.1.3, and its about that
timeframe that I noticed the change in how Networker recovers files from
multiple savesets.  Its my suspicion that the upgrade is what caused the
change in behavior.

Its added a lot of time to our recovery time, since in order to do a
recover now, it has to demux /u01 from tape, then when done, demux /u02
from tape, and so on, even when /u01 and /u02 were muxed together during
backup.

Is there any way to get Networker 7.1.3 to recover like I remember
Networker 6.1.3?  

I hope that what I'm asking makes sense.  It could be that I am
remembering wrong and Networker has always recovered savesets serially,
but I don't think so, because I used to recover ~150 Gbyte from 6
savesets (full backups) and it took ~3 hours.  Now its taking about 5x
longer because it does one at a time, even if two or more of the
savesets to be recovered are muxed together on tape.


-- 
Craig Ruefenacht
UNIX Systems Administrator
USANA Health Sciences
http://www.usana.com

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listserv.temple DOT edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
also view and post messages to the list. Questions regarding this list
should be sent to stan AT temple DOT edu
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

<Prev in Thread] Current Thread [Next in Thread>