From: Alan B. <aj...@ms...> - 2005-05-09 15:20:38
|
On Mon, 9 May 2005, Jeffrey L. Taylor wrote: > I am new to Bacula so consider this an idea to look at, not a > definitive answer to do. If this were my problem, I'd look at the > retention of jobs. If the last full backup is purged before doing the > last incremental in the week, I would expect the incremental to be > upgraded to a full backup, because there is no longer any record of > a full backup. The problem is not that the backup got upgraded to a full or differential one. The problem is that the full or differential backup happens in the incremental pool - and if separate pools are used, the incremental pool is usually tiny by comparison to the full one. As an example, one of my backup sets has 6 terabytes allocated per full backup cycle, but only a few hundred Mb for incrementals because the data is mostly static/appending star observation image sets , rather than highly dynamic items like databases. AB |