In response to Jeff Tucker <[EMAIL PROTECTED]>: > > We'd like easy access to recent files for the case where someone deletes > the wrong file. We also want to take full backups (probably monthly) > offsite for long-term archiving and disaster recovery. > > My idea right now is to use a second machine with a TB of disk storage > as a near-line backup machine. I would do a monthly full backup and then > 30 days of incremental backups to disk. At the same time as the monthly > full disk backup, I could do a monthly full tape backup. That tape would > be taken offsite. > > Does this make sense? It seems like the disk backups would be more > convenient and faster to manage than rotating pools of tapes. Any other > ideas?
We do this here. The disk storage is nice and fast compared to tapes. We do weekly offsites, though, so a little more intensive than what you're doing, and a single full backup is ~150G. > On the disk storage server, how would I go about setting up my volumes? I would recommend a full/differential/incremental rotation (which is pretty typical) Create a monthly pool, weekly pool, and daily pool. Have fulls got to monthly, differentials to weekly, incrementals to daily. Retention times: monthly 2 months, weekly, 4 weeks, daily, 7 days. The reason you want to keep at least 2 monthly backups is that if you have it shorter than that, and your full backup fails for some reason, it will have partially overwritten (i.e. corrupted) your only full backup! By having at least two full backup sets at any time, you avoid this. The differentials reduce the amount of storage space you need for backups between fulls, but the reduced the amount of resolution you have for restoring, so you need to consider your requirements carefully. The layout I describe is pretty typical. > I think I basically want a full backup with a volume retention period of > 25 days and incremental backup volumes with retention periods of 35 days > or so. So, if I do a full backup every month, it will overwrite the old > full backup. Yes. Again this creates a problem if the backup fails for some reason -- you lose you previous full backup. > And, if I do daily incrementals, then I will have a > rotating pool of 35 dailys. Yes. > Do the old dailys get deleted automatically? > Do they just get overwritten? The latter. After the retention period expires, the volume is marked "purged", but no data is destroyed. When the system needs a new volume, it take the oldest purged volume and truncates it and begins anew. > When I start out, using automatic volume > naming, will the system count up to 35, then start overwriting the older > volumes? Pretty much. The first month you'll see new volumes created every day, the second (and subsequent) month you'll see the system reuse them. > If someone could give me an idea what these volumes or pools should look > like in the config files, I'd appreciate it. Or, if I'm totally off > base, let me know. I believe the layout I describe above is documented in the Bacula docs, if not, it's typical enough that I'm sure there are examples out there. Just schedule two full jobs for the beginning of the month: one to tape and one to disk. Do the tape job first so future incrementals are based off the disk job. You could also do job cloning, but I don't think that's necessary to get what you want. -- Bill Moran Collaborative Fusion Inc. ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
