GFS with multiple, rotating external drives?


Author
Message
CarbonatedTeal
CarbonatedTeal
New Member
New Member (2 reputation)New Member (2 reputation)New Member (2 reputation)New Member (2 reputation)New Member (2 reputation)New Member (2 reputation)New Member (2 reputation)New Member (2 reputation)New Member (2 reputation)
Group: Forum Members
Posts: 1, Visits: 3

Hi all.

I’ve been using Macrium for ages to backup my workstation, and I think I’m doing it all wrong.

My current setup is like this:

1) I have two external drives, each that mount on a different drive letter.
2) I have two backup jobs created, Job A, for the drive letter of Drive A, and Job B, for the drive letter of Drive B.
3) Week 1, I manually run a full backup on Drive A using Job A, and switch it for Drive B.
4) Week 2, I manually run a full backup on Drive B using Job B, and switch it for Drive A.
5) Week 3, I manually run an incremental on Drive A using Job A, and switch it for Drive B.
6) Week 4, I manually run an incremental on Drive B using Job B, and switch it for drive A.

Rinse and repeat. Roughly monthly, at the time of a drive swap, I’ll delete all the backups on the incoming drive and restart with a full backup.

….There has to be a better way to do this.

My goal would be to keep separate GFS backups on each drive, so that I can do a full restore from each drive without needing the other. I’d like to be able to start the backup manually (click “run next” or somesuch rather than having to schedule it, because the workstation isn’t always on, but I’m willing to compromise on that.)

And now the monkey wrench – I have some folders that are also backed up using Crashplan. Does Crashplan mess with the “Archive” flag? Will this affect my backups?

Does running this as an image backup vs a file/folder backup have different considerations?

Does running the “wrong” job manually (per the retention rules of the job) result in the “right” job being automatically run? For example, if I set the job to retain four incrementals and then accidentally tell it to run a fifth, will a full be run instead?

Can anyone help me set this up in a way that’s saner?

Thanks in advance all!


Edited 1 November 2018 11:22 PM by CarbonatedTeal
jphughan
jphughan
Macrium Evangelist
Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)Macrium Evangelist (6.1K reputation)
Group: Forum Members
Posts: 4.2K, Visits: 30K
Quite a bit to unpack here, so settle in, and maybe grab some coffee!  I have a disk rotation involving 9 destination disks and everything runs off a single definition file that contains a single schedule, so yes there are easier ways to manage this.  But I'll start with the easier stuff first.  Here goes:

- It's not clear from your post or the section of the forum you posted in whether you're performing image jobs or File & Folder jobs, but either way, CrashPlan will not have any effect on Reflect.  The Archive flag doesn't come into play with either type of job.  So that's not a monkey wrench at all.

- If you have a retention policy to keep 4 Incrementals, then when you run the 5th, the oldest Incremental will be consolidated out.  The retention policy doesn't mean that Reflect blocks you from creating a certain type of backup past that point.  It just means that when you create one more of those backups, the oldest one gets deleted -- or consolidated in the case of Incrementals, since a straight deletion of an Incremental would invalidate all subsequent backups in the chain.  How exactly this consolidation works depends on whether you have a Differential in your set, which I can't tell because you mentioned using a GFS strategy, which implies Differentials, but you didn't mention any use of Differentials in the routine you described.  If you only use Fulls and Incrementals, that would be an FS (Father-Son) strategy.  Anyhow, if you have a Differential somewhere in your set, then the oldest Incremental and second oldest Incremental would be consolidated, which essentially deletes the oldest Incremental as an available restore point without breaking the Incremental chain.  Macrium has a nice animation of this concept here. If you don't have a Differential and also don't have Synthetic Fulls enabled, the same thing happens.  If you don't have a Differential in your set and do have Synthetic Fulls enabled, then the oldest Incremental and the Full would be consolidated.  Macrium has a nice animation of this concept here.

- If you have backups going to different disks, then each one already has "self-contained" backups, i.e. you will never need both disks to perform a restore -- so you're already covered there.  But in terms of simplifying your setup, it sounds like you don't have both disks connected at the same time, correct?  That's good from a safety perspective, and it also means you can simplify things.  If the only reason you currently have separate jobs is because you have two destinations, but they are otherwise configured identically, then right off the bat you can eliminate one of those jobs and then modify the remaining job to use the "Alternative locations" feature.  If you edit your definition file, you'll see a link for that directly under the destination field.  Add the other drive's target folder there, and then whenever you run that one job, Reflect will send the backup to whichever of those locations is available at the time.

- In terms of backup strategy, as you may have noticed, one of the issues around disk rotations when using traditional configurations, especially if you ever want to use scheduling, is that when your strategy involves multiple backup types (Full, Diff, Inc), then you have to worry about making sure the right disk is connected on "Full day", how long it's been since that specific disk got a Full, etc.  The easiest way to sidestep all of that is to switch to the Incrementals Forever with Synthetic Fulls strategy.  The reason is that under that strategy, the only job you ever run is an Incremental, so you never have to worry about which disk is connected at which time in order to get the "right" backup type onto it.  This backup strategy is shown in the second link I posted above.  Note that since you'll only ever have one Full per disk under this strategy, you may want to increase your Incremental retention policy so that you retain the desired amount of history on each disk -- but Incrementals Forever with Synthetic Fulls is the most storage-efficient way to store a given backup history anyway, precisely because there's only one Full on the disk (and no Diffs), which minimizes data duplication.  If you only had one target disk, that would arguably make this setup a risk (because if the Full was corrupt or partly unreadable, all of your backups become useless), but having a disk rotation means you'll always have two completely independent backup sets, so you're covered there.

- In terms of scheduling vs. manual runs, there's no reason you can't keep doing manual backups, or you could switch to schedules, or you could do both.  Manual backups can be executed at any time by right-clicking your definition file, clicking Run Now, and selecting the desired backup type.  You can also choose to create a desktop shortcut to run the backup for convenience.  And you can keep doing that whenever you want even if you have a schedule active.  But I'm not sure what your "Run Next" vision is.  It sounds like you're already running a manual backup, so what are you envisioning there?

Phew!  Well,I realize that was probably a lot to take in, so take some time to digest, and let me know if you have any further questions. As I said, I use a disk rotation strategy myself, and I'm happy to help. Smile

GO

Merge Selected

Merge into selected topic...



Merge into merge target...



Merge into a specific topic ID...




Similar Topics

Reading This Topic

Login

Explore
Messages
Mentions
Search