I've come accross a few problems with Unity servers where the SQL transaction log for the UnityDB grows until it fills the disk and starts causing problems (i.e. 7gb+).
As I have some background in SQL it's easy enough to fix... but I've seen this post:
It states we need Full recovery model, or the 'nightly truncates' will fail. If they do fail, surely this can be ignored as truncations are not required with the Simple recover model?
Anyone any other reason full recovery would be required? The difference AFAIK is that with simple, the transaction log is cycled and therefore kept small.
With full, it is written to, and can grow, and the log is only truncated/reused after a full backup.
If, therefore, our customers don't back up their DB regularly (and we all know they don't!) they get this problem... eventually it grows.
Why would full recovery be required when even if we did a full system restore we wouldn't likely do a transaction log restore, as we have a full DB backup taken at the point the whole system was backed up?