We have just seen this issue on an exchange server, check out the below links for more info.
Concurrent backups of the same database are not allowed in Exchange 2010. There can be only one single backup job running against a given database irrespective of whether the backups are done with the Store Writer or the Replication Writer. This is achieved by Exchange store putting the database in a “backup-in-progress” state, this in-memory state is cleared either at the completion of the backup process or when the service is restarted. When it is used, the Replication Writer communicates with the active database owned by the store process to verify that there is no active backup in progress against the selected database, and then sets the “backup-in-progress” state for the selected databases. Restarting services that host Exchange Writers, complete OS reboots as well as cluster failovers will result in losing the in-memory “backup-in-progress” state together with its associated data, which in turn will cause the backup job to fail.
If there are no backups overlapping this is what I would do, and why:
1.- Stop NetWorker services on Exchange servers, including Replication Manager Exchange Interface and Replication Manager AgentPS.
2.- Delete /nsr/tmp folder from Exchange servers.
3.- Start NetWorker services including Replication Manager Exchange Interface, but leave Replication Manager AgentPS service stopped, as it's a manual service and will be started up when required during a backup operation.
Then run this command and ensure that all writers are in status [1] no error:
vssamdin list writers
If all writers are ok, then run a new backup (remember that backup in DAG will be ran against ONLY one of the nodes).
If you are running a backup against the 2 physical nodes then that is the problem, as Exchange 2010 automatically marks the DB's as "Backup in progress" when there is a backup operation ongoing.
Mabro1
2 Intern
•
666 Posts
0
February 14th, 2012 02:00
This discussion has been Moved to the Networker Support Forum to ensure your question gets to the right audience.
Thanks and regards,
Mark
5N7PCR1Jdd12579
1 Rookie
•
2 Posts
0
February 14th, 2012 12:00
Hi Mike
We have just seen this issue on an exchange server, check out the below links for more info.
Concurrent backups of the same database are not allowed in Exchange 2010. There can be only one single backup job running against a given database irrespective of whether the backups are done with the Store Writer or the Replication Writer. This is achieved by Exchange store putting the database in a “backup-in-progress” state, this in-memory state is cleared either at the completion of the backup process or when the service is restarted. When it is used, the Replication Writer communicates with the active database owned by the store process to verify that there is no active backup in progress against the selected database, and then sets the “backup-in-progress” state for the selected databases. Restarting services that host Exchange Writers, complete OS reboots as well as cluster failovers will result in losing the in-memory “backup-in-progress” state together with its associated data, which in turn will cause the backup job to fail.
http://msdn.microsoft.com/en-us/library/bb204080(v=exchg.140).aspx
http://aspoc.net/wp-content/uploads/2010/05/EMC_NetWorker_Exchange_2010_Support.pdf
CarlosRojas
1.7K Posts
0
February 14th, 2012 21:00
Hi Mike,
If there are no backups overlapping this is what I would do, and why:
1.- Stop NetWorker services on Exchange servers, including Replication Manager Exchange Interface and Replication Manager AgentPS.
2.- Delete /nsr/tmp folder from Exchange servers.
3.- Start NetWorker services including Replication Manager Exchange Interface, but leave Replication Manager AgentPS service stopped, as it's a manual service and will be started up when required during a backup operation.
Then run this command and ensure that all writers are in status [1] no error:
vssamdin list writers
If all writers are ok, then run a new backup (remember that backup in DAG will be ran against ONLY one of the nodes).
If you are running a backup against the 2 physical nodes then that is the problem, as Exchange 2010 automatically marks the DB's as "Backup in progress" when there is a backup operation ongoing.
Thank you.
Carlos.