Hi all, my current customer has an exchange 2010 DAG cluster consisting of 10 servers, organized in pairs. Each database in the DAG is passive on one host and active on another. Because they want to restrict the size of each individual Exchange database there are now 210 exchange databases, each not larger than 500 GB. Because everything is one DAG, I organized the back-ups to be owned by one exchange DAG node. This is more and more becoming a problem. A ’tdpexcc q TSM * FULL /all’ takes well over 45 minutes. Restoring a single database from full and incremental back-ups takes 2 hours to query TSM and exchange…. a bit much for our patience.
And of course, in inventory expiration this one DAG node is the bottleneck as well. I turns out that a VSS backup of a single database is not just one backup object, but 10’s of thousands. For this one node about 40M objects in total… Inventory expiration for the DAG node takes more time that for all other nodes (about 400) combined. Is there a reason why I must stick to the one DAG node? Or can I just as well create a DAG node for each pair of exchange servers in the DAG? This will eventually reduce the number of objects per node by a factor of 5, making things a lot more manageable. -- Met vriendelijke groeten/Kind Regards, Remco Post r.p...@plcs.nl +31 6 248 21 622