Does anybody have experience with "bigger" DP8 or 9 Internal Databases and how to tweak them for peek efficiency?
Due to Exchange 2013 I have to merge two older Cells under one new DP9 but Performance seen so far in Tests is way worse than the old Raima DB.
Target Cell will then have aprox. 450 Clients, 50 LTO Drives, 50 virtual LTO Drives and 6 FileLibraries/StoreOnce-Devices. Sessions per Day should be around 1200 at the beginning, expecting 1.5k within half a year.
With all DCBFs migrated we currently estimate 150GB of IDB.
Customer expects Extended Checks to be done in less then 6h, current Test-Env needs almost 5h for one check, and that is only a 15GB IDB.
Also the Response-Times for Restore-Selection have to be rapid fast, right now it already takes 2min just to expand any item in the Restore-Tree, imagine you have to select several dozend Folders, Files or Exchange Databases.
Any hints on what to consider in such an situation?
Better go towards faster Disks for better IO or more Cores for more parallelism or less but faster cores for all the single-threaded tasks?
Something Special to consider with the AppServer or the pgbouncer maybe?
Thanks for any suggestion