• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 2007
  • Last Modified:

Will re-enabling a long-inactive DFS replication member cause problems?

Several years ago we set up a three-way DFS replication between servers at different sites (two are Server 2008 R2, one is SBS2008).  Because of lack of bandwidth at one site, we disabled the replication to and from that site, leaving the remaining two (which have been and are still replicating fine).

We now have improved bandwidth at all sites, so want to re-enable replication.  Should it simply be a matter of re-enabling the connections, or can we expect problems because of the time that has passed?

The folder that is replicated has been used by all three sites, although each site has its own sub-folder and doesn't usually make changes to the other folders.  We have compared the files and folders and there are no conflicting files (same name with different contents) but many orphans (exist only on one side).

If there's any chance of (for example) a folder deletion from years ago being synchronised, then how can we prevent this?  We want to be certain that all current files remain and are consolidated.

Thanks in advance!
0
David Haycox
Asked:
David Haycox
  • 5
  • 5
  • 3
  • +2
2 Solutions
 
compdigit44Commented:
I was thinking you could do a one-way replication marking one of the folders as read-only .. Full article is listed here but has draw backs..

http://blogs.technet.com/b/filecab/archive/2007/08/16/using-one-way-connections-in-dfs-replication.aspx

Another option is to do a once time robocopy
0
 
kevinhsiehCommented:
You would stand a good chance of having your files overwritten.

I would move that folder outside of the replication group hierarchy on the filesystem and create a new replication group for just that folder. You will need to adjust the share path and any DFS paths.

I have been using DFS Replication for about 6 years. I think I use it basically the same way you do. It's best to try to keep files that get modified on separate servers in different replication groups.  One reason why is that just this past week I found that a server had a dirty shutdown of DFS Replication which causes replication to stop for that member and when you run the command to restart it all changes made on that member will be lost as the copies on the other members are considered authoritative. What I ended up doing was deleting the replication group and recreated it so that I could specify the specific server that had the correct version of the files. I knew that I didn't need.to worry about what was on the other servers because I only allow one server to actively share a set of files at a time. This prevents conflicts when the same file is being modified at the same time on multiple servers.
0
 
kevinhsiehCommented:
I believe that one way connections are deprecated and I don't trust that you could mark the other members as being read only and have replication work the way you expect until replication is caught up. In addition, you can't mark subfolders as read only, only the entire replication folder.
0
Free Backup Tool for VMware and Hyper-V

Restore full virtual machine or individual guest files from 19 common file systems directly from the backup file. Schedule VM backups with PowerShell scripts. Set desired time, lean back and let the script to notify you via email upon completion.  

 
MaheshArchitectCommented:
After disabling replication If you have any new data stored on this member \ modified existing data, just copy \ backup that data before you again enable replication.
Because after a disabled member is enabled, the member must complete an initial replication \ sync of the replicated folder. any updated or new files present on the member will be moved to the DfsrPrivate\PreExisting folder on the member, and it will be replaced with authoritative files from another member.
Once initial sync gets completed, you can copy previous data again to member and it will get replicated to other replication partner

The membership changes must be replicated to all domain controllers, and the member must poll active directory to obtain the changes.

You need to force AD replication (may be run repadmin /syncall) on domain controllers and also run dfsrdiag pollad command on member to pickup changes
0
 
David HaycoxAuthor Commented:
Mahesh, that's pretty much what I was hoping and in fact I've been backing up the files (at each site, for good measure) over the weekend.

So just to be clear, you would expect the previously disabled member to "reset" to match the primary member, after which I can just restore the files from backup?

How would I know when I would be able to start restoring the files?
0
 
MaheshArchitectCommented:
Yes, once initial sync is completed by member, check for Event ID: 4104 Source: DFSR on member server under dfsr logs
Also one more testing is you can go to DFSR management console \ replication group \ connection objects and manually right click connection objects and select "Replicate now"
Once initial sync is completed, it should successfully replicate the connections

Once replication is successful, means initial replication is completed and then you can restore new \ modified files to member, it should automatically get replicated to another member
0
 
David HaycoxAuthor Commented:
Just enabled the connections and ran "replicate now" - the following error popped up:

The member SERVERNAME could not be contacted to get current schedule status.  The error returned was "The DFS Replication WMI connection or synchronization object was not found."
Do you still want to attempt to override the schedule for this member?

Connectivity between the sites is working fine.  I assume if I agree to override the schedule, it won't make any difference because of this error?

Ah, perhaps I was just impatient after running the replication commands suggested above.  After waiting a short while it appears to be working.  So far so good...
0
 
kevinhsiehCommented:
You are going to replicate all data twice. Once to push the old data onto your server, and again after you restore the good data. In addition, any time a server stops replicating (think server crash or power outage) you risk having to backup and restore files, and you can't enable access until files.have replicated and then been restored. IMHO that's crazy. Break up the replication groups so that the active set of files on one server are not in the same replication group as the active files on another server. I have several TB of replicated files and it would be literally impossible to to do what you're trying to do and keep my business running. A full resync of one server takes over a week it seems.
0
 
David HaycoxAuthor Commented:
Hmm, I take your point.  By separating the replication groups for the different folders you can have more granular control over the whole process.  Of course, to change it to that now would require replicating everything again, right?  Including the site which is already up-to-date because it's been connected all along?  I'll definitely bear that in mind in future.

In this case there's nothing like as much data as you have (about 300GB) so I'll leave it running and see what happens - for now.
0
 
kevinhsiehCommented:
Your current process requires replicating everything anyway after you restore. If you move all of the files the replication engine should still be able to compare the existing files to the updates, assuming you have Windows 2008 R2 Enterprise or better on one side or Windows 2012 or better on one side of the connection.

Yes, more granular control is better, along with smaller groups and volumes. I have learned all this the hard way over many years.
0
 
David HaycoxAuthor Commented:
So far so good, nothing deleted.  Security permissions on one folder got reset to how they were on the long-disabled version, but that's all (so far).  It's replicating happily from the re-enabled site back to the primary member (still a few days to go by the looks of it) but as yet nothing has copied the other way, although thousands of files are showing as backlogged.  Is this expected behaviour or should it be going both ways at the same time?  I have told it to "replicate now".
0
 
kevinhsiehCommented:
I don't think you will get two way replication until it is fully resynced.
0
 
MaheshArchitectCommented:
That's right, as mentioned in my earlier comment and Kevin last comment, the remote site server will not replicate any data to partner unless initial sync get completed
Check Initial Sync topic in below articles to understand Initial Sync
https://technet.microsoft.com/en-in/library/dn495052.aspx
http://blogs.technet.com/b/askds/archive/2008/02/12/get-out-and-push-getting-the-most-out-of-dfsr-pre-staging.aspx
0
 
David HaycoxAuthor Commented:
So far so good, 75GB of 110GB has been replicated one way.  Once that's done we'll see what happens with the other direction.
0
 
Seth SimmonsSr. Systems AdministratorCommented:
I've requested that this question be deleted for the following reason:

Not enough information to confirm an answer.
0

Featured Post

[Webinar] Cloud and Mobile-First Strategy

Maybe you’ve fully adopted the cloud since the beginning. Or maybe you started with on-prem resources but are pursuing a “cloud and mobile first” strategy. Getting to that end state has its challenges. Discover how to build out a 100% cloud and mobile IT strategy in this webinar.

  • 5
  • 5
  • 3
  • +2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now