11-06-2017 05:09 AM
Sanity check.
When running USL tasks should also be moved to the new server right?
Solved! Go to Solution.
11-06-2017 11:56 PM
Hello all,
Highlevel. 2 Datacenters, each having Exchange Journal Dag's and EV servers. 1 DC is in DR (running in the other DC), with the servers being shutdown...
I 'resolved' this issue by reconfiguring EV. The tasks are Journal Tasks, which are not recreated during USL apparently. I have verified MSMQ and Storagequeue on the servers which were in DR were empty. Then removed the Journal Mailboxes from the targeted Exchange servers (in EV), and added them to the Exchange servers which were still funtioning. That works. It does mean that when we fail back, the same excercise needs to be done, to restore the 'standard' configuration.
I am yet to find out the order in which the DR was done. It might have been possible to change the system mailbox on the tasks for the Exchange servers going into DR to the system mailbox of the server where the databases were going to, but that requires investigation in our acceptance environment... The Exchange which were shut down are up again, due to issues with the logfile replication..
Thanks.
GJ
11-06-2017 05:45 AM
@GertjanA Yes the tasks are recreated on the new server automagically. However, I always run the following SQL query to clear the MSMQ before starting the task again.
USE EnterpriseVaultDirectory
UPDATE StorageServiceEntry
SET StorageArchive = ' ', StorageRestore = ' ', StorageReplayIndex = ' ', StorageSpool = ' ' Where ComputerEntryID = 'your old server name'
UPDATE RetrievalTask
SET RetrievalSpoolQueue = ' ' Where ComputerEntryID = 'your old server name'
UPDATE ArchivingRetrievalTask
SET MessageQueue = ' ' Where ComputerEntryID = 'your old server name'
UPDATE RetrievalTask
SET MessageQueue = ' ' Where ComputerEntryID = 'your old server name'
UPDATE JournalTask
SET MessageQueue = ' ' Where ComputerEntryID = 'your old server name'
UPDATE PublicFolderTask
SET MessageQueue = ' ' Where ComputerEntryID = 'your old server name'
11-06-2017 11:56 PM
Hello all,
Highlevel. 2 Datacenters, each having Exchange Journal Dag's and EV servers. 1 DC is in DR (running in the other DC), with the servers being shutdown...
I 'resolved' this issue by reconfiguring EV. The tasks are Journal Tasks, which are not recreated during USL apparently. I have verified MSMQ and Storagequeue on the servers which were in DR were empty. Then removed the Journal Mailboxes from the targeted Exchange servers (in EV), and added them to the Exchange servers which were still funtioning. That works. It does mean that when we fail back, the same excercise needs to be done, to restore the 'standard' configuration.
I am yet to find out the order in which the DR was done. It might have been possible to change the system mailbox on the tasks for the Exchange servers going into DR to the system mailbox of the server where the databases were going to, but that requires investigation in our acceptance environment... The Exchange which were shut down are up again, due to issues with the logfile replication..
Thanks.
GJ
11-07-2017 12:16 AM
That's great!
So to clarify, you shutdown Exchange AND EV in primary Data Center and moved both Exchange AND EV to the DR (now active) Data Center? Ofcourse yes, that would require the tasks to be recreated, as the Journal targets are not on the same Exchange Journaling Servers (although they are on the same DAG databases) and would require the System Mailboxes from the DR Exchange Journaling Servers to be specified.
Within one Data Center or even across Data Centers, if Exchange servers are left running, and only USL is used for EV, then the tasks are recreated. Is this correct?
11-07-2017 01:28 AM
Hello Arnold,
I'm not sure about the tasks being recreated. I recall from memory Mailbox Archiving tasks are, but I am not sure about Journal tasks... Will attempt a test once we have our acceptance environment up and running.
As for the DR, it is unknown yet HOW it has been performed. The advise was to first do EV, and make sure all was ok. We now believe that DR for both EV and Exchange was done in parallel, probably with skipping of some checks. The joys of having a remote organization doing most of the work. But, lesson learned for the future. This will give me headaches after the return to normal, as that will require not only performing the same excercise, but also on items left in msmq and storagequeues. The workinstruction does have 'run all tasks in reportmode for a few hours prior to action' but you never know. Oh well, keeps me busy.