[mythtv-users] LVM / data questions

[Hika van den Hoven]

Hoi Anthony,

Friday, July 4, 2014, 10:34:11 AM, you wrote:

> On 4 July 2014 16:45, Matt Boyd <mattslists at gmail.com> wrote:
>  
> Hi mythtv people,
> I've a big bucket of data that I'd like to store somewhere and the
> mythtv mailing list is about the only place I frequent where this
> sort of thing has been discussed. 
>  
>  Bascially, I've several terabytes of data across different usb
> drives. It's copies of technical and other data I use with
> reasonable metadata so I can work out what it is when I need it. I
> don't need to access it all regularly but it's not exactly in a
> useful directory structure and there's too much of it for me to even
> think about setting up something like that. I've been using windows
> search and a few other resources to hunt down the data I need when I
> need it which is not optimal but it does work.
>   
> The data isn't so important to me that I can afford to have an
> enterprise level backup system for it.

> Ideally, what I would be able to do would be to split the data
> across a number of usb hard drives linked into a really redundant
> file system so that if one drive went down I could put another in,
> recover what I could from it and deal with (or not) any corrupted
> files when I noticed them. I'd also like the filesystem to be easily
> readable if the drives were pulled out and accessed individually.
>   
> This is about the level of redundancy I'd like from my mythtv video data as well.

> So maybe I'm looking for something that just disperses my data
> across whatever pile of disks I give it access to. I don't need the
> intricacies of a raid setup, but I need something more automated
> than just manually working out how to split directories and index them across drives.
>   
> Does anyone have any thoughts? I've started to look at LVM but it
> doesn't seem to have robustness that I want.

> Thanks

> Matt
>  
>  
> Take a look at mdadm RAID

> https://raid.wiki.kernel.org/index.php/RAID_setup

> Anthony

lvm2 is not about redundancy, it's about flexibility. You can easily
resize, create or delete volumes on the fly. I use it on my server on
top of raid1.

If it's about redundancy, as Anthony says use raid1. It would work on
usb, but you have to make sure the drives come up together if you
attach them at runtime. Also you have to create a deactivation script
for disconnecting at runtime. Better for mdraid to have the drives
permanently attached. I've only used degraded arrays (just one of the
two drives) on usb to get data off old arrays.

Basically what raid1 does is to keep two (or more) copies of your
filesystem in a way transparent to all users of the data. For every
array you need two drives or partitions equal in size. If one is
bigger the extra size is not used.
support for raid is build in the kernel. You however need the mdadmin
pakkage to manage and if you want monitor the arrays.


Tot mails,
  Hika                            mailto:hikavdh at gmail.com

"Zonder hoop kun je niet leven
Zonder leven is er geen hoop
Het eeuwige dilemma
Zeker als je hoop moet vernietigen om te kunnen overleven!"

De lerende Mens