Friday, April 20, 2012 4:35 AM
Hi Microsoft Forum Team,
I am gonna implement DFS Replication. The replication will be done from One VM to Physical Server.
i am using windows server 2008 R2 for DFS.
- Please tell me is there any issue you have faced related on this ? It would be helpful for me to be causious.
- please tel me what are the prerequisites need to be follow before implementing.
- Please tell me what are the Do's and Dont's on DFS.
- i am going to use only data replication to other sever. Do i need to configure Namespace for this ?
- which protocol does it use for DFS and please tell me the port number
Monday, April 23, 2012 6:53 AM
Monday, April 23, 2012 7:14 AMModerator
1. It's OK to sync files between VM and physical servers by using DFSR, however, there is some known issue for VMs that may prevent DFSR working as expected after VMs restore from a snapshot, please refer to:
2. Please make sure that the firewall between DFSR servers will not block DFSR network traffic.
3. I suggest you to read the following articles before you start with DFSR:
DFS Step-by-step Guide For Window Server 2008: http://technet.microsoft.com/en-us/library/cc732863(v=WS.10).aspx
4. No, you don't.
5.DFSR syncs files through MS-RPC protocol, as it's well known, MS-RPC is using dynamic RPC port ranges from 1025 to 65535. But you can set a static port for DFSR by using dfsrdiag staticRPC command, please follow:Regards,Cicely
- Marked As Answer by Cicely FengMicrosoft Contingent Staff, Moderator Thursday, May 03, 2012 12:52 AM
Wednesday, April 25, 2012 5:36 AM
thanks for the reply. one more query comes up.
i am going to configure two way replication between the VM and physical Server for redundancy.
if suppose my Physical Server is down. and after some days it went UP. can the VM server able to replicate the data automatically from the day in which VM server is down? please let me know.
Thursday, April 26, 2012 8:55 AMModerator
Yes, after the server is up again, the replication will continue from the point of where it stopped. However, you may not want to let the server down for a long time as there would be lots of backlog for the server after it ups again.
Besides, if the DC is down for a long time, be cautious about problems may caused by lingering objects. Refer to the link about how to fix it if needed: http://technet.microsoft.com/en-us/library/cc738018(v=WS.10).aspx
Thursday, April 26, 2012 10:20 AM
thanks for the reply. Please find my below queries.
1. Shall i disable to Replication process once the any one server is down? if we do so any issue behind this?
2. i have 400 GB data to be replicated for the first time. After i configured the replication for both the servers, how long wil it take to replicate the data ? does it replicate everything without any issue in one go? any bandwidth issue will come ? is that advisable to start replicate in the production time..
both the servers are location in a same rack with 1 Gbps connection
Friday, April 27, 2012 7:29 AMModerator
1. You don't need to do anything for replication once any of the server is down.2. It's hard to say the accurate time it will take to replicate data, I think, because DFS Repliaction throttles bandwidth by throttling RPC calls, various buffers in lower levels of the network may interfere and cause bursts of network traffic.
There is one article above, DFS Rplication FAQ, mentioned this, please check it for details: http://technet.microsoft.com/en-us/library/cc773238(v=WS.10).aspx#BKMK_022
For boosting the initial sync for DFSR, prestage(pre-seeding) data for both servers before DFSR initial replication is a good option for you if it will take a long time to complete the replication , which will save lots of time.
Share two articles with you:
Designing Distributed File Systems: http://technet.microsoft.com/en-us/library/cc772778(v=WS.10).aspx
Replacing DFSR Member Hardware or OS (Part 2: Pre-seeding): http://blogs.technet.com/b/askds/archive/2010/09/07/replacing-dfsr-member-hardware-or-os-part-2-pre-seeding.aspx
Thursday, May 03, 2012 4:07 AM
thanks for the brief explaination.
we are nearly going to implement this. i will get back to this thread if i have any doubts or issues..