Great $ Subject is working with a 14 node cluster and writing with an awesome speed to a CoFC on a SO 5100,
No single execution of a huge job works flawless
There are no 2 run's with an identical behavior
To master it, we're gonna split it (256 VM's and 70 TB of data) up into different Jobs and present the 3PAR vcopy's to different backup servers. But even at our initial splitting we still encountered "non-predictable & non-repeatible errors"
Bottom line Q: how to stabilise this kind of backups?
cause when all is virtualized and centralised with one huge consolidation, our backup should be able to cope with it ALL WAYS.....
or do we start pointing to Microsoft?