4 ms·
On a lark a made a Ceph cluster a couple of years ago with a dozen RPis, a dozen commodity 1TB USB drives, and a SO/HO switch and... it worked. Like, it was a j
by bandrami 3y ago
On a lark a made a Ceph cluster a couple of years ago with a dozen RPis, a dozen commodity 1TB USB drives, and a SO/HO switch and... it worked. Like, it was a joke project one afternoon but suddenly we had a 3X redundant 4TB S3 pool that was faster than our WAN connection and just took switching out a HD or RPi once a quarter or so whenever Nagios lit up that one had failed.
- api 3y agoYep, but companies need to pay tens or hundreds of thousands a month for AWS to run it for them. I feel like we are ripe for another turn of the on-prem/off-prem wheel. Then, of course, once it's all back on-prem we will once again have price gouging monopolies (probably in software) and someone will have the bright idea of moving everything to... what will they call it next time? Maybe "the grid" instead of "the cloud?"