mediahound
Dabbler
- Joined
- Mar 11, 2013
- Messages
- 15
Because my head is still stretched a little beyond capacity trying to understand this new way of dealing with drive arrays, volumes, and everything else, I still don't know if I understand all the terminology or what is what, so bear with me.
I can set up ZFS on a single external volume too, and have it do all of the necessary integrity verification, scrubbing, and similar, correct? Like if everything I needed fit on a single 4tb drive, I could connect it, mount it, and while accessing the files have it scrub the entire drive to make sure everything is intact?
From what I understand I could also do this with a group of drives even of mismatched sizes? Say I have a mix of 3tb and 4tb external USB drives, and I create a pool from them, under the assumption that all the drives will always be connected as a group every time. (used similar to Just A Bunch Of Disks storage, not mirrors or striped performance arrays) Ie say 14tb total storage, but lets say because I know how modern drives lose bits of data, i'd like it to use 1tb or so for parity (or the equivalent of parity information to recover bad data discovered during scrubbing) or whatever other size I want... ie I assume I dont have to set it to the size of an entire drive, or designate a single dedicated drive to be all parity like with a RAID array, I can just manually specify however much I think is reasonable? (fully realizing if I don't set up enough, and I get more corruption than that between the last scrub I lose it all - i'm aware of this) I might even have two separate volumes on the group of drives, high priority data (which has 2 copies of everything to insure it will get through) and lower priority data (which has 1 copy plus some parity info to reconstruct partial bit rot or bad sector loss) without having to pre-define the size of each yet.
What i'm wanting to do is to create volumes of relatively known total size (most of them larger than a single external hard disk but whenever a cluster of disks is used it would always use them all together) to be used for migrating data in (ie - someone sends me a group of 3 drives with however it has to be set up to recognize of data to import), migrating data out (ie - I send a mirror of certain data I have, say 6 drives, for a sneakernet backup to someone else to do the same on their array), while maintaining ZFS's "end to end integrity verification", metadata, and every time the drives are connected (they would not be often connected spending time in the mail) forcing a total scrubbing of ALL data to insure integrity across the stored volume so that it's not degrading. Once that's done the entire block of drives would be disconnected. (either because it was the import array, export array, or backup array of drives put on the shelf to check again in a month or two's time)
From what I understand and have been reading so far I should be able to do all this, correct? (I don't need exact answers yet, like exact instructions how, just an expert opinion of approximately how this is normally done or what commands/terms I might want to read up more on and understand better before finalizing plans if it doesn't work the way I think it does)
I can set up ZFS on a single external volume too, and have it do all of the necessary integrity verification, scrubbing, and similar, correct? Like if everything I needed fit on a single 4tb drive, I could connect it, mount it, and while accessing the files have it scrub the entire drive to make sure everything is intact?
From what I understand I could also do this with a group of drives even of mismatched sizes? Say I have a mix of 3tb and 4tb external USB drives, and I create a pool from them, under the assumption that all the drives will always be connected as a group every time. (used similar to Just A Bunch Of Disks storage, not mirrors or striped performance arrays) Ie say 14tb total storage, but lets say because I know how modern drives lose bits of data, i'd like it to use 1tb or so for parity (or the equivalent of parity information to recover bad data discovered during scrubbing) or whatever other size I want... ie I assume I dont have to set it to the size of an entire drive, or designate a single dedicated drive to be all parity like with a RAID array, I can just manually specify however much I think is reasonable? (fully realizing if I don't set up enough, and I get more corruption than that between the last scrub I lose it all - i'm aware of this) I might even have two separate volumes on the group of drives, high priority data (which has 2 copies of everything to insure it will get through) and lower priority data (which has 1 copy plus some parity info to reconstruct partial bit rot or bad sector loss) without having to pre-define the size of each yet.
What i'm wanting to do is to create volumes of relatively known total size (most of them larger than a single external hard disk but whenever a cluster of disks is used it would always use them all together) to be used for migrating data in (ie - someone sends me a group of 3 drives with however it has to be set up to recognize of data to import), migrating data out (ie - I send a mirror of certain data I have, say 6 drives, for a sneakernet backup to someone else to do the same on their array), while maintaining ZFS's "end to end integrity verification", metadata, and every time the drives are connected (they would not be often connected spending time in the mail) forcing a total scrubbing of ALL data to insure integrity across the stored volume so that it's not degrading. Once that's done the entire block of drives would be disconnected. (either because it was the import array, export array, or backup array of drives put on the shelf to check again in a month or two's time)
From what I understand and have been reading so far I should be able to do all this, correct? (I don't need exact answers yet, like exact instructions how, just an expert opinion of approximately how this is normally done or what commands/terms I might want to read up more on and understand better before finalizing plans if it doesn't work the way I think it does)
