SnapImage is designed specifically for this problem. It does a copy-on-write
soft snapshot of the volume. Then it slurps the filetable and parses it
directly to extract the file retrieval metadata. Then it does a raw image
backup that does not care about file granularity. That means that for
millions of files it is much faster than file-by-file backup.
It also does block-level incrementals which works pretty well because it
uses the snapshot driver to track the modified blocks and build the
incremental block map in real time, so when incrementals run it already
knows what to back up as soon as it starts.
The volume does have to be a basic disk though.
Pluses - it gives significantly faster backup times compared to standard
NetWorker when a volume contains millions of files. It provides browsable
file recovery with comparable performance to regular NetWorker on a similar
filesystem. It provides block level incremental backup with the same
benefits.
Negatives - it uses a filesystem level write-intercept driver. This means it
is sensitive to OS revisions and other third-party drivers. It uses NDMP as
the control mechanism (all required licensing included) which means it
doesn't work exactly the same as your other backups. For example you can't
use scanner to rebuild file indexes for SnapImage backups. You have to
designate a dedicated write-copy-cache volume on the server that is being
backed up.
It has enough extra things to manage that you don't want to use it unless
your other approaches can't solve the problem. But when the pain of the
problem is worse than the extra management required for SnapImage, it's a
good fit.
-----Original Message-----
From: Kobus Strydom [mailto:kstrydom AT MEDIHELP.CO DOT ZA]
Sent: Thursday, October 26, 2006 6:51 PM
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Incremental backup of large volume takes longer than full
Please can someone help.
I have a Windows 2003 server HPDL580 4 x 3.2Ghz, 4GB ram, 72GB local
disk, 500GB EMC volume, Gigabit ether, 2GB Fiber for backup purpose.
Networker 7.3.1
On the EMC volume are very roughly about 30mil files, all images.
A full backup of this volume takes roughly 36hours. There are about
10GB's worth of new images added daily, yet the incrementals that I
perform takes forever and seldom completes before 11:am the next working
day. They start at 18:00 each night. The speed I get out of incrementals
are in the region of 500KB to 1MB p/s.
I have journal manager enabled on this volume, yet it looks lik all
files are checked for changes?
Any suggestions/help would be greatly appreciated. I have worked with
our local supplier on this issue for months now and still no joy. They
suggested a raw backup, which is great, I get an average speed of 50MB
p/s, but this doesn't help when I need to restore individual files.
Kobus Strydom
Systems Engineer
Medihelp
Mobile +27 82 9090882
Tel +27 12 3342295
Fax +27 12 3342426
e-mail kstrydom AT medihelp.co DOT za
Medihelp is an Authorised Financial Services Provider
---------------------------------------------------------------------------
NOTICE: Please note that this eMail, and the contents thereof,
is subject to the standard Medihelp eMail disclaimer which may be found at:
http://www.medihelp.co.za/disclaimer.htm
<<<< GWAVASIG >>>>
To sign off this list, send email to listserv AT listserv.temple DOT edu and
type
"signoff networker" in the
body of the email. Please write to networker-request AT listserv.temple DOT edu
if
you have any problems
wit this list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
To sign off this list, send email to listserv AT listserv.temple DOT edu and
type "signoff networker" in the
body of the email. Please write to networker-request AT listserv.temple DOT edu
if you have any problems
wit this list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|
