We have used this to cut down on large incrementals.  I remember the default
settings for the the change journal are too small for large drives/volumes
and would not be used/overfill on some drives.  These were the 2 settings;
"% of volume for log file" and "% of log for Allocation Delta" in the change
journal manager.   I know we changed the setting from the defaults but
forget how much we increased them.

I know the change journal and these setting must be manually set for each
drive?   After changing these settings and activating the change journal a
full backup is needed to clear it out.  We found just turning on the change
journal did not work without a full backup.

That said, the change journal cut down on large incrementals. The
performance appears to be better but that may be because its backing up less
changes/data.   I say that because I had expected better performance with
the change journal; one would think it could consult this quickly rather
than tranversing directories looking for changes/archive bits.   I may, in
fact, have the same probelm as you do.    Anybody else getting good
performance with the change journal?

Robert Maiello
Thomson Healthcare


On Wed, 2 Jun 2004 16:04:58 +0200, Voetelink D. <voetelink AT ECN DOT NL> wrote:

>Hello networker friends,
>
>I'm looking at the NTFS change journal to see if it is a possible
>solution for the very long time an incremental backup on our
>windows 2000 servers takes, due to the virusscanner.
>
>It seems that there's not much helpful information on the subject.
>
>I've set up a test system, containing a disk with a lot of (sub)directories,
>and even more small files. When I do an incremental without the change journal
>enabled, it takes about 20 minutes to complete. If I do an incremental backup
>with the change journal enabled, it takes 5 minutes, even though only one
>small file has been changed. After some testing and monitoring it seems that
>it's taking 5 minutes just to 'read' the change journal.
>
>Also, it seems that sometimes the change journal isn't used or perhaps
disabled,
>without any obvious cause. Just like the change journal gets activated
>sometimes without any cause (we have it disabled everywhere since we know so
>little about it).
>
>The only useful tool I found to read the changejournal is 'fsutil',
supplied with XP.
>
>
>Some questions I have:
>- How long will the reading of the change journal take if the system is 'in
use'?
>  will this always be 5 minutes? Or will this be dependant on the size of the
>  journal?
>
>- How can I be sure that all changes are in the journal? The output of
fsutil doesn't seem
>  to include the date/time of the entry. If I understand correctly, the
journal is a FIFO,
>  so if it's full, the oldest entries get kicked out.
>  So how can I ensure that all changes are still in the journal at the end
of the day?
>  (or a week later at the level1 backup?).
>
>
>Any thoughts/insights anybody? Any experience?
>
>
>greetings and thanks,
>
>Dennis
>
>--
>Note: To sign off this list, send a "signoff networker" command via email
>to listserv AT listmail.temple DOT edu or visit the list's Web site at
>http://listmail.temple.edu/archives/networker.html where you can
>also view and post messages to the list.
>=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

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