Hello again,
If I'm correct in my research:
NTFS Change Journal (http://technet.microsoft.com/en-us/library/
cc781134.aspx#w2k3tr_ntfs_how_nxcy)
As files, folders, and other NTFS objects are added, deleted, and
modified, NTFS enters change journal records in streams, one for each
volume on the computer.
The total size of all the records currently in the journal varies, but
there is a configurable maximum size. The change journal can exceed
the maximum size until the size reaches an outer threshold, at which
point a portion of the oldest records are deleted until the change
journal is restored to its maximum size. The maximum size of the
change journal is configurable but cannot be reduced, only increased.
The change journal conveys significant scalability benefits to
applications that might otherwise need to scan an entire volume for
changes. File system indexing, replication managers, virus scanners,
and incremental backup applications can benefit from using the change
journal.
The change journal is much more efficient than time stamps or file
notifications for determining changes in a particular namespace.
Applications that must rescan an entire volume to determine changes
can now scan once and subsequently refer to the change journal. The I/
O cost depends on how many files have changed, not on how many files
exist on the volume.
The APIs are fully documented and can be leveraged by independent
software vendors (ISVs). Microsoft uses the change journal in Windows
Server 2003 components such as the Indexing Service and File
Replication Service. ISVs can use this feature to enhance the
scalability and robustness of a range of products including backup,
antivirus, and auditing tools.
(http://technet.microsoft.com/en-us/library/bb742450.aspx)
The Change Journal is initially an empty file on the disk volume. As
changes occur to the volume, records are appended to the end of this
file. Each record is assigned a 64-bit identifier called an Update
Sequence Number (USN). When Microsoft was first developing the Change
Journal, it was internally called the USN Journal. That's why the
structures and defines in the winioctl.h header file refer to the
Change Journal as the USN Journal. When a record is added to the
journal, it is assigned a USN. USNs are generated in increasing order,
so that you can compare USNs to find out the order of events (lower
USNs are older events). USNs are not contiguous, so it's possible that
the first USN record might be 0 and the second USN record might be
128.
Is it safe to say that DFS-R keeps a DB of various things, including,
when the latest USN entries; so that it can hop right into the Change
Journal and review changes made since then?
Post by Matt Brown - nycThanks for replying.
Beyond the RDC function, I am concerned with how it "surveys the
land." If the DFSR service is down, how is the USN journal watched
for changes? How does DFSR know which files were changes?
I suppose I'm more interested in the interactions of DFSR and the DFS
Jet database? I don't really want to narrow down the answers,
though... any input is appreciated.
Thanks,
Matt Brown
On Mar 13, 10:29 am, "Isaac Oben [MCITP,MCSE]"
Post by Isaac Oben [MCITP,MCSE]Hello Matt,
DFS-R uses differential compression which allows only changes to a file to
be sent over the network. If the service goes down and then starts again, it
will start comparing based in diff compression and it will determine the
most recent changes.
--
Isaac Oben [MCTIP:EA, MCSE]
Post by Matt Brown - nycHello,
In theoretical scenario where the DFSR service is down for an extended
period of time, and a high number of file changes (specifically
changes), and a lower number of new files are created, how does DFS
perform the reconciliation process?
This article's bullet-points are a good introduction to the entire
process (http://technet.microsoft.com/en-us/library/cc781091.aspx),
except that it specifically states: "DFS Replication detects changes
on the volume by monitoring the update sequence number (USN)
journal." What if DFSR is down when these changes are made?
Does the DFSR service go out and survey all of the files for changes?
Thanks!
Matt Brown