When I try to copy a 14GB file from a Server 2008 server to a Server 2008 R2 server, it copies about 5GB worth and then returns a Network Error "There is a problem accessing filename Make sure you are connected to the network and try again". I can copy the file from the 2008 server to a 2003 R2 server with no problems, I can also copy a large file from a server 2008 R2 box to another R2 box. The problem only seems to happen between an earlier version of Server and Server 2008 R2. One thing I didn notice that while copying a large file between to server 2008 R2 servers that the transfer rate was slower than what I was getting between the 2008 to 2008 R2 before the copy failed, so I'm not sure if there is some factor there. I found some threads relating to large file copies between XP and Vista when Vista was 1st released, but nothing for Server 2008 R2

Hello MrFig, From your description, it appears you encounter the copy scenario as followed: When you copy a large file from Windows Server 2008 to a Windows Server 2008 R2, there will be return a Network Error "There is a problem accessing filename Make sure you are connected to the network and try again." Based on my test, we cannot reproduce the issue in our local environment. We can copy a large file with 12GB in size from Windows Server 2008 to Windows Server 2008 R2 without problem. To troubleshoot the issue on your side, I would like to suggest you first obtain and install Windows Server 2008 Service Pack 2 onto your problematic server. How to obtain the latest service pack for Windows Server 2008 http://support.microsoft.com/kb/968849/ After applying it, please restart the server and then test it again with copying a file in a stable network environment to see if the Error will be re-occurs. Meanwhile, please try copy methods with both copying via Windows Explorer and copying via command line (ex. Robocopy or Copy) Does this issue both exists in both of the above copy methods? Best regards, David ShenThis posting is provided "AS IS" with no warranties, and confers no rights.

The server in question does have Service Pack 2 installed. Also, just for the sake of curiosity... I was able to copy the file from the Server 2008 server to a Server 2003 R2 server without a problem. I then tried to copy the file from the Serer 2003 R2 server to the Server 2008 R2 server and received the same error.I did try copying the file using both Windows Explorer and Robocopy - Same ResultI then setup FTP on the original box and then FTP'ed the file in questionto my local workstation (which happens to be a Server 2008 R2 box). I was then able to copy the file from my workstation to the desired destination via Windows Explorer.Here are some more details about our environment if it is helpful:Source Server(s):- Server 2008 Datacenter with Service Pack 2- Source disk is a iSCSI target- Using Microsoft iSCSI initiatorDestination Server:- Server 2008 R2 Datacenter with all updates from Windows Update installed- Member of a 5 node Hyper-V Cluster- Destination disk is a iSCSI target configured as a Clustered Shared Volume- Using Microsoft iSCSI initiatorThanks,Lamar

Additional information:I did some more testing and found that the problem is not a general problem between server 2008 and server 2008 R2 because I was able copy the file in question from the source to my workstation without any problem (again my workstation is runngin server 2008 R2 with desktop experience turned on).After doing several tests the problem is specifically between any prior version of windows (I tested server 2008 and server 2003 R2) copying to the iSCSI target of our Hyper-V cluster. (note: The copy of the filefrom the 2003 R2 server was located on local attached disk).....***Update***Well, I'm going to have to scratch what I have been saying. As I have been writing this, I was running some additional tests between my workstation and the Cluster and it also failed this time. The new commonality that I'm am seeing is in the transfer speed. The copies that were successful were plodding along at a consistent 10-15 MB/second. The copies that failed the speed had bursts up in the 70-100 MB/second range and would copy at that rate for several seconds then pause for several seconds and sometime resume again for another short burst then pause again and then fail as I have been mentioning.Right now it appears that the copy process is successful if the speed stays below 20 MB/Second. Note: even in this case the copy process seems to have momentary pauses, but the process does eventually resume and complete.So I'm thinking that the problem is some kind of networking related issue where some buffer/cache is getting full and the timeout is coming into play while the buffer or cache is clearing out or something.Any thoughts?Thanks,Lamar

Hi Lamar, You can try test with disabling TCP chimney offload and RSS on Windows Server 2008, and then check if the issue still exist. To disable TCP Chimney Offload, follow these steps: netsh int tcp set global chimney=disabled To disable RSS, follow these steps: netsh int tcp set global RSS=disabled To disable 'Receive Window Auto-Tuning Level', follow these steps: netsh interface tcp set global autotuninglevel=disabled Hope it helps.This posting is provided "AS IS" with no warranties, and confers no rights.

Hi Lamar,I want to see if the information provided was helpful. Please keep us posted on your progress and let us know if you have any additional questions or concerns. Best Regards, David ShenThis posting is provided "AS IS" with no warranties, and confers no rights.

Hi David,Thanks for the input. Sorry for the delay, but I have been out for a few days.I tried making the three changes that you suggested on one of the nodes of the cluster (the one that currently owns the storage) and then tried copying the file again. From that node, I copied the file from the source to the destination. The results are as follows:- for the first few seconds the file transfered at a rate of about 90 MB/Second- then the copy process paused for what I would say was a good 20 seconds- The copy process resumed at about 12 MB/Second- Later the process dropped down to about 7-9MB/Second and completed the copySo whatever is going on, it appears that changing those setting helped (of course I have only made one attempt thus far), however, I'm not sure that we would want to keep those setting disabled long term.Any thoughts?Thanks

Hi MrFig, TCP Chimney offload is a new feature which helps to offload the network throughput when CPU is in high utilization. You can keep them disabled if there isn't any network issue in long team. However, if the CPU of the node always in high utilization, you may consider re-enabling it. For your more reference: Information about the TCP Chimney Offload, Receive Side Scaling, and Network Direct Memory Access features in Windows Server 2008 http://support.microsoft.com/kb/951037 The effect of TCP Chimney offload on viewing network traffic http://blogs.technet.com/networking/archive/2008/11/14/the-effect-of-tcp-chimney-offload-on-viewing-network-traffic.aspx Hope it helps.This posting is provided "AS IS" with no warranties, and confers no rights.

Hi David,We found a solution that seems to have fixed the problem. We noticed that when we installed Hyper-V and the network adapters were virtualized that the "Link-Layer Topology Discovery Mapper I/O *" options on the physical network adapters were unchecked. When we re-enabled the Link-Layer Topologies everything seemed to work much better.Thoughts?Thanks,Lamar

This Microsoft hotfix may help: http://support.microsoft.com/kb/2477730 Roger Faucher Roger Faucher Printing and Image MVP - Since 2004