Geek: Backup Strategies and 1.4TB on the way
I just plunked down $400 on two 750GB drives. Plus $20 for another drive enclosure. Meaning soon, up to 1.5TB of storage in a fit-in-my notebag portable fashion, and will be all the dataI own. Pretty amazing.
My current backup strategy is segmenting the types of data
CRITICAL (about 3MB every minute, 200MB changed in a given day) ->
1GB flash drive
jungledrive.com -> Amazon S3
Eclipse and SVN
DISK IMAGES -> windows backup + scheduler -> external harddrive.
Currently I own about 1.1TB of storage (about 65% filled), mostly on one machine:
320GB SATA
320GB SATA
250GB SATA
120GB SATA
40GB IDE (old)
With all those drives it's not surprising the power supply started to fail.
One my work machine, I have everything critical in one folder called data_sync, includes projects, my documents, palm. Rather than 20 places, as many apps like to do when installed. Of this only 2MB is ultra critical, it's the todo and idea capture lists. On average about 30-150 items change a day.
For the largest space offenders, movies (all the DVD's I own are about 750GB). I realized that Netflix and Itunes/ebay is a reasonable backup strategy, in addition to just getting rid of them. I don't watch TV, and it's rare I watch movies, and projecting forward I don't anticipate that changing: Every year I get like 5 books I never have time to finish reading. Then again when I watch I do it on a projector, and HD DVD is likely to be better.
Home server motherboard died trying to upgrade the power supply. I could have gotten a replacement motherboard on ebay ($30-$60) but I did a Total Quality of Life calculation, and Total Cost of Ownership and found that it wasn't worth it.
- the cost of maintaining another machine outweighs the benefits at this stage. 2 machines is redundancy, synchronizing 3 machines is 6 possible interdependencies. Drives, updates, licenses etc.
- having the drive embedded in the server outweighs the utility of having frequently when I travel.
- 1TB drives are still too pricey, at $400/each or so.
one pain in the butt is the raid controller on the server doesn't write standard drives, so I can't just pop out the existing drives and put them in drive enclosures, as I would like.
It took some time to analyze, and rationally I really would have rather put the money toward product development, but it's constant brain suck not having access to the data, worrying about resynching gigabytes. Given the wide variety of options (e.g. different drive combos, motherboards etc), the opportunity costs, it really overloads my optimal solution checking (aka 'perfection paralysis', and puts me in a bad mood. So at some point one has to measure the amount of time spent thinking on this subjects rather than thinking about something else more productive and add that in.
My current backup strategy is segmenting the types of data
CRITICAL (about 3MB every minute, 200MB changed in a given day) ->
1GB flash drive
jungledrive.com -> Amazon S3
Eclipse and SVN
DISK IMAGES -> windows backup + scheduler -> external harddrive.
Currently I own about 1.1TB of storage (about 65% filled), mostly on one machine:
320GB SATA
320GB SATA
250GB SATA
120GB SATA
40GB IDE (old)
With all those drives it's not surprising the power supply started to fail.
One my work machine, I have everything critical in one folder called data_sync, includes projects, my documents, palm. Rather than 20 places, as many apps like to do when installed. Of this only 2MB is ultra critical, it's the todo and idea capture lists. On average about 30-150 items change a day.
For the largest space offenders, movies (all the DVD's I own are about 750GB). I realized that Netflix and Itunes/ebay is a reasonable backup strategy, in addition to just getting rid of them. I don't watch TV, and it's rare I watch movies, and projecting forward I don't anticipate that changing: Every year I get like 5 books I never have time to finish reading. Then again when I watch I do it on a projector, and HD DVD is likely to be better.
Home server motherboard died trying to upgrade the power supply. I could have gotten a replacement motherboard on ebay ($30-$60) but I did a Total Quality of Life calculation, and Total Cost of Ownership and found that it wasn't worth it.
- the cost of maintaining another machine outweighs the benefits at this stage. 2 machines is redundancy, synchronizing 3 machines is 6 possible interdependencies. Drives, updates, licenses etc.
- having the drive embedded in the server outweighs the utility of having frequently when I travel.
- 1TB drives are still too pricey, at $400/each or so.
one pain in the butt is the raid controller on the server doesn't write standard drives, so I can't just pop out the existing drives and put them in drive enclosures, as I would like.
It took some time to analyze, and rationally I really would have rather put the money toward product development, but it's constant brain suck not having access to the data, worrying about resynching gigabytes. Given the wide variety of options (e.g. different drive combos, motherboards etc), the opportunity costs, it really overloads my optimal solution checking (aka 'perfection paralysis', and puts me in a bad mood. So at some point one has to measure the amount of time spent thinking on this subjects rather than thinking about something else more productive and add that in.