Data Storage & Backup

Working On It...

Thinking...

Data starts off flat, and we impose convenience structure. Should it always be available locally? Is it irreplacable? Who's the artist? Who likes to consume it? Those are orthoganal dimensions, but all usefull. Filesystems are (without linking) capable of just a single organizaion, so which one wins? I would hope all the orgazational dimension could be realized in some way (e.g. music file metadata and capable browsers), but if we get just one (which is still better than zero (flat (everything under a single directory)), seems like it should be whichever serves the machine, since user facing tools (eg media players) are smarter than the backend stuff (for now).

I'd think the most important, and of concern to the backend is the data's lifecycle. There's stuff that we always want available (locally), stuff that's just a copy of that stuff, and stuff the we don't need around for long or don't mind of it expires. In other words, an archive, backups, and cache.

archive

The archive is mostly irreplaceable (eg photos, personal webiste, finances, etc) plus purchased digital only artifacts (eg downloaded music).

What about music that we have discs for, or video we have media for? By "disk is cheap" we should have it in the archive too since we want in "online" (on demand), so just think as the original media as another backup (but not even a great one because it deosn't include metadata we've added).

Content organization doesn't matter too much because things will change over the decades the archive will be changing. For now I'm thinking of directory chunking by dates of major additions.

backup

We obviously backup the archive. Hosts (computers) are backup up too (though not archived).

Quartly snapshots two years back (8 rdiff-backup sessions)?

cache

Intend to consume once, like video clips, podcasts, movie rentals, broadcast recordings. Maybe some move to the archive. If we loose it we don't care much, so it's not backed up.

Namespace

Thinking...

The namespace is presented by the fileserver and reflects common usage patterns. Things that come tomind...

• archive - date named directories of collections as they come in
• music - "Artist - Album" (not necessarily)
• video, new - a cache
• video, kid
• video, educational
• video, home - personal
• movie, new - a cache
• movie, library - stuff we want to keep around
• book
• exercise - yoga
• audio - podcast, episode
• type, video - for apps that need it; links to directories of video, movie, home movie, exercise, etc
• backup?

• archive
• music
• kid
• dave
• jeanie
• home video

• /srv/vol/1 for tv, video rental & library - not mirrored or backed-up
• /srv/vol/2 for online music, archive, and video personal
• /media/backup for backup of online music, archive, and video personal

3 drives >= 1.5T ?

• online - 2,500 (2)
• backup - 1,250 (1)
• travel - ?

Drive Models

• 2TB Samsung Spinpoint F4 HD204UI 5400 32MB - $100
• 1.5TB Seagate Barracuda 7200.11 ST31500341AS 7200 32MB - pooly reviewd at newegg, $60
• 2TB Seagate Barracuda LP ST32000542AS 5900 32MB - pooly reviewd at newegg, $70
• 1.5TB WD Caviar Green WD15EARS 64MB - pooly reviewd at newegg (doa, "advanced formatting (4k sectors)), $60
• 2TB WD Caviar Green WD20EADS 32MB - pooly reviewd at newegg (doa)), $120

Needs

Storage

Usage/Backup

Inventory