We will be moving fromExchange 2003 to Exchange 2007 very shortly. We have2 mailbox servers with Windows 2008, running Exchange 2007 SP1, Update Rollup 5, Enterprise edition. We have created the mailstores on each server with the location name preceeding the Mailstore name (so names are unique), and both servers have all mailstores on their D drive, with consistent naming for them. One Mail server "ServerMX4" is in Melbourne, the other "ServerMX3" is in Adelaide. Latency is 15ms and our slowest link is 6Mbps. ServerMX4 has 800 mailboxes and 450Gb, and ServerMX3 has 1000 mailboxes and 700Gb.I tested that SCR replication works, and there are no errors, all directories were successfully created on the target. But, we have worked out that we will not be able to transfer the entire database on each server across the Internet, as it will take months and saturate our link. I asked about throttling SCRhttp://social.technet.microsoft.com/Forums/en-US/exchangesvrdeploy/thread/3b63d1c7-b00f-4310-b565-5ec20bfc06c3, but have not had any help yet.I have consultedhttp://technet.microsoft.com/en-us/library/bb738131.aspxabout suspending SCR, resuming SCR, and seeding the target database. But, one thing that is not clear - does the source database need to be off-line for the entire time that the data is being copied?We intend to dismount a mailstore, copy all files to a USB drive or backup tape (as a file, not Exchange backup), and thencourier the USBdrive across to theother[target] office, and restore the file into the SCR target folder on the "StandbyMachine".This is so that the source database is only offline for the time it takes to copy data locally across a fast link, not the time it takes to copy across the Internet. We intend to re-mount the "source" database as soon as the copy is finished, and then it will be a day or two until the USB drive arrives in the "StandbyMachine" target site.Will there be problems if the pre-seeded database file on the "StandbyMachine" is older than the current database on the SCR "source"? Once we re-enable SCR, will the existing EDB copy that we have done be valid, or will all data have to be sent across the wire?Finally, for the copy of thesource database to USB drive, would wecopy theEDB files and the Log files too?(our log files are on a seperate RAID set, seperate RAID controller) Any advice would be great!

A few weks ago I performed the following test steps:SCR-Source: Suspend-StorageGroupCopy Dismount-Database Eseutil /mh SCR-Target: Stop MSExchangeRepl service. I did this because I do not know if the service is happy is somebody copies the database without using Exchange aware commands. It might be the case that Resume-StorageGroupCopy tells the service to check what has changed in the meantime. Delete database, log files, checkpoint file Copy offline database from SCR-Source. I did not copy any log file and the checkpoint file because if you dismount a database then you will have a clean shutdown and do not need the logs. Start MSExchangeRepl service. I had to start it twice because the first attempt failed with an timeout error. SCR-Source: Mount-database Resume-StoragegroupCopy Get-StorageGroupCopyStatus shows initializing Use LoadGen to generate new logs on the SCR-Source. After some time the SCR-Target copied all log files from the SCR-Source. This includes also the log files that had already been applied to the database when I dismounted the database on the SCR-Source.

Thanks for the step-by-step instructions, but it still does not answer my question;If I dismountthe mailstore "SurnamesD-E" on "serverMX3" and then copy it to a temporary USB drive, remount the mailstore "SurnamesD-E" on the machine"serverMX3" - people start using the mailstore"SurnamesD-E", then three days later (allowing for courier delays) copy the data from the USB drive to "serverMX4", and then re-activate SCR for this Storage Group.Will SCR try to replicate the entireStorage Group"SurnamesD-E" (that I just copied) again? Will it try to transmit the whole 100Gb over the Internet or only the50Mb ofchanges since I took the off-line copy?If I do my usual backups of the Storage Group, will this remove logs and then break SCR?I have had to suspend SCR on all my servers until I get this sorted out - a single mailbox of 1.2Gb was replicating and caused problems for more of our sensitive applications.

Hi Christian, Firstly, some basic information: When enabling an existing storage group for SCR, a manual seeding of the storage group may be necessary. If the transaction log file containing the createDB record (usually the log with the first generation sequence number) no longer exists, you must manually seed the SCR target after enabling SCR. Therefore, if the log with the first generation sequence number does not exist, the automatically seeding will not occur. Therefore, the SCR will not replicate the entire storage group again. It will just copy the transaction log file in the transaction log path on the source server to the target server. Therefore, you can apply the steps provided by J-H to enable SCR. You can test the situation firstly before applying it to your product environment. Regarding Transaction log truncation after enabling SCR, you can refer to following article: Please review the section Log Truncation http://technet.microsoft.com/hi-in/library/cc535020(en-us).aspx#DbResScen Mike

I think it will only copy the log files and not the database file.Please refer to the link Mike provided. Logs are not truncated on the source if they have not been inspected on the SCR target.