TFS Integration Platform Error while Syncing - Caught and converted, The request was aborted: The request was canceled

Question

Hello,
We are doing a 2-way version control sync between TFS 2013 servers. 
TFS A - Project A is getting synched with TFS B - Project A.
TFS A - Project A has got all the code and files, whereas TFS B - Project A is empty.
Previously we have synced TFS 2013 projects across servers with out any error or issues. Today we came across an issue (check log at end of post) and hope Microsoft can
provide some guidance on this. Following is the problem we are facing today:
We have configured the TFS Integration platform tool for VC - 2 way sync and started the migration. Current
status is that the discovery, analysis is complete and migration is half way when I started to get such errors
(check log at end of post) .
Note that the migration has not stopped...
After doing some research and googling I understand that the reason for this error is that by default if
an individual file transfer cannot complete 16MB in 5 minutes, the operation will fail. And the solution to this is to modify
chunk size in the file of "TfsMigrationShell.exe.config". I have few questions regarding this...
Question 1:Is
above understanding and solution correct? I also notice that this error appears for every huge file (approx >30MB), is there any timeout setting at the target server? Is this something to be changed at the Tool level config or the Target IIS level
config?
Question 2: Suppose
I do not change the config setting, then huge files are skipped due to this error. However, after the migration of the other files is finished, then, will the tool retry to sync the skipped (huge) files? Or has it skipped them forever?
Question 3:
Suppose I modify
chunk size in the file of "TfsMigrationShell.exe.config", then, which of the following do I need to do (give me the best solution among the 3 below)
a) Do I need
to restart the migration (by deleting and recreating the project on the 2nd TFS Server and re-starting migration)? 
b) Do
I need to create a new configuration and restart migration?
c) Do I need to restart the sync service?
The concern here is that we have got 3 more jobs (TFS
Integration tool configurations) running parallelly that
is syncing Work Items & VC of other TFS projects and want to make sure that restarting the TFS Integration Job Service will not have any negative impact on the other jobs as well as no negative impact on the current migration job...
Microsoft, Please advice. Thanks!
Log:
[1/21/2015 10:55:39 AM] AsyncUploadFileState.Completed -- caught and converted:
[1/21/2015 10:55:39 AM] Microsoft.TeamFoundation.VersionControl.Client.VersionControlException: C:\TfsIPData\10\ications\MDM Folder\Binaries\v9.5\DMS Web Service Release 1.5.zip: The request was aborted: The request was canceled. ---> System.Net.WebException: The request was aborted: The request was canceled. ---> System.IO.IOException: Cannot close stream until all bytes are written.
[1/21/2015 10:55:39 AM] at System.Net.ConnectStream.CloseInternal(Boolean internalCall, Boolean aborting)
[1/21/2015 10:55:39 AM] --- End of inner exception stack trace ---
[1/21/2015 10:55:39 AM] at System.Net.ConnectStream.CloseInternal(Boolean internalCall, Boolean aborting)
[1/21/2015 10:55:39 AM] at System.Net.ConnectStream.System.Net.ICloseEx.CloseEx(CloseExState closeState)
[1/21/2015 10:55:39 AM] at System.Net.ConnectStream.Dispose(Boolean disposing)
[1/21/2015 10:55:39 AM] at System.IO.Stream.Close()
[1/21/2015 10:55:39 AM] at System.IO.Stream.Dispose()
[1/21/2015 10:55:39 AM] at Microsoft.TeamFoundation.VersionControl.Client.FileUploader.UploadChunk(FileStream fileContentStream, Int64 end)
[1/21/2015 10:55:39 AM] --- End of inner exception stack trace ---
[1/21/2015 10:55:39 AM] at Microsoft.TeamFoundation.VersionControl.Client.Workspace.EndUploadFile(IAsyncResult asyncResult, PendingChange& pendingChange, Byte[]& uploadedContentBaselineFileGuid, Int64& uncompressedLength)
[1/21/2015 10:55:39 AM] at Microsoft.TeamFoundation.VersionControl.Client.Workspace.AsyncUploadFileState.Completed(IAsyncResult asyncResult)

Answer 1

Hi Charles,
I have been doing in depth research and testing of this and here are my findings:
Replies to your points:
1) It failed with UploadChunkSize value change. Tried 1MB, 500KB, even 100KB. Fails with same error! Is there any other setting we can change for the timeouts at the target TFS server also?
2) The huge files are not skipped. The migration loops on the failed files and doesn't continue to process the next lot of files. So the migration is blocked on the huge files.
3) OK. Thanks. 
Please see below the finding on when migration succeeds and when it fails:-
Scenario 1:
a) Where Migration fails:
TFS A is in India and TFS B is in UK. Both TFS are on different domain. Migration tool is installed on same machine as TFS A. There is a firewall that allows the access from TFS A machine to TFS B machine. The migration tool fails to work in this scenario
when it encounters a file size approx > 8MB. I tried setting the UploadChunkSize value to 100KB. Yet it failed!
b) Where Migration works:
However, if the files are approx less than 3MB then the migration works without any errors.
Scenario 2 where TFS migration tool works:
TFS A is in India and TFS B is also in India (located closeby). Both TFS are on different domain. Migration tool is installed on same machine as TFS A. There is no firewall in this case. No changes were made to UploadChunkSize value and files with size
of 10's and 100's of MBs got synced without any issue.
Questions
Question 1: Can we conclude now that there is some issue with the firewall or network in case of scenario 1? As mentioned above Scenario 1 works if file size is less then approx 3MB.
Question 2: Is there any timeout setting on the Target TFS server which can be set so that it waits for the file without timing out?
Thanks for any help...