Tuesday, October 19, 2010

Renewing Edge Synchronization

If you have added any domain on HUB server and you are not able to see this domain on Edge server then might be your synchronization between your hub and edge servers are not working. To test this you can run following command;

Test-EdgeSynchronization

If the connection result is showing failed then you need to do edge subscription again for this edge server. Following are the steps for this

1. On the edge server which is not synchronizing run following command;

New-EdgeSubscription -filename "C:\temp\edgesub.xml"

now copy this .xml file to hub server and run following command

New-EdgeSubscription -filename "C:\temp\edgesub.xml" –site London

Please note the site parameter is required to bind this edge server to a specific AD site.

Now run following command

Start-EdgeSynchronization

You will notice that now the edge server is synchronizing data with HUB server.

For verification you can run test command again and this time it should show you the success message

Test-EdgeSynchronization

Monday, October 18, 2010

Enabling SCR on WAN

SCR is an Exchange 2007 DR solution available from Exchange 2007 SP1 onwards.  You can enable SCR on LAN and also on high latency networks. SCR also gives you benefit to recover from a Datacenter failover scenario. Since last month i was working to enable SCR between our DCs connected via MPLS. After a thorough analysis and lots of try and error i have compiled following results;

SCR Enabling Steps:

1. Seeding of .Edb File.

2. Enable Replication.

3. Monitor Replication

 

Step 1: Seeding:

Seeding is a prerequisite for enabling SCR, we can enable seeding by automatic way and manual way, because i have to seed .edb file on MPLS which is a high latency network i have to go for manual procedure. Following steps i have followed for manual seeding.

1. Take a snapshot of storage group on which i want to enable SCR (we are using snapshot based backup solution) you can also take the normal backup of the storage group. This will also truncate all the committed logs and we have less log files to copy from source to destination after enabling replication.

2. Run  below Command

Enable-StorageGroupCopy -Identity StorageGroup1 -StandbyMachine London1 -ReplayLagTime 0.1:0:0 –SeedingPostponed

The above command will enable SCR on storage group “StorageGroup1”. Standby machine (destination server) used is London and the replay lag time i have mentioned is 1 minute. seeding postponed is used because i don't want to enable seeding automatically i want to copy .edb file manually to destination server.

After running the above command, log files for this storage group will not truncate unless until we enable replication. This means that the drive space of log files folder should also be carefully planned because if the latency is high and the .edb file is taking days to copy then the log file drive space might going to be fill which will affect the production environment.

3. Take a snapshot on the storage group again or take a normal storage group backup again.

4. Ask the backup team to provide the .edb file from the last snapshot.

5. Use TeraCopy to copy .edb file from source to destination server. make sure the drive letters for log and data drive are same on standby server because it is one of the prerequisites of SCR.

2. Enabling Replication:

Replication will copy all logs files from source to destination. Replication is an ongoing process and whenever a log file is created on source it has to be shipped to destination for making both source and destination mailbox databases in same state.

Following is the command for enabling replication

Resume-StorageGroupCopy -Identity StorageGroup1 -StandbyMachine London1

after running the above command you will see that the log files started to copy from source to destination server. Standby server will replay these log files after 50 logs files has been copied. The first log file after you enable SCR should be present on destination otherwise replication will fail.

3. Monitoring Replication:

It is very important to check the status of storage group copy. The below command will show us the health of SCR and how many files in copy queue and replay queue.

get-storagegroupcopystatus StorageGroup1 -standbymachine London1

If you are seeing failed  in the summarycopystatus field then there is a possibility that the database you have seeded and the logs files copied are not at the same state (LSN number mismatch) or might be you don't have the first log file available to replay on the destination.

If you are seeing large number of files in copy queue length then bandwidth might be the possible cause.

You can also run following command on standby server for checking replication status;

Test-ReplicationHealth