Archives of the TeradataForum
Message Posted: Mon, 01 Jul 2002 @ 19:39:29 GMT
| Subj: || || Re: Cluster archives vs. database level archives |
| From: || || Christopher Platt |
The biggest benefit I see from Cluster archives is breaking your box up into equal parts so that ALL backup jobs run in about the same
period of time, depending on data skew, takes the unpredictability out of when your jobs will finish. You never know if data has been
heavily loaded into a table from week to week which can cause a job to take longer then it previously did. If you use clusters you can
eliminate this headache as the extra data from this load is spread out over (x) number of jobs. This also helps to take away some of the
headaches if new tables are added if you code an archive (dbc) all, exclude (dbc) it will sweep the box from week to week.
Restoring can be a bit tedious as you have to run a restore job for each cluster that you back up and then a bld job . So if you split
your box into 4-5 clusters it will take 5-6 jobs to restore one table.