Many use just HDDs. The load delay is usually because they don’t want to spin up every disk just for one chunk of a file, so they collect jobs and schedule spin up and down times.
But there are dedicated services that indeed use tape.
Many use just HDDs. The load delay is usually because they don’t want to spin up every disk just for one chunk of a file, so they collect jobs and schedule spin up and down times.
But there are dedicated services that indeed use tape.
Depending on what cloud tier you are using they might use tape then (e.g. IBM Cloud Object Storage Deep Archive).
It is marked as a breaking change in the release notes though.
I have two WD Red 3TB with 109.858h and a Samsung SSD 830 256GB with 110.639h on my main workstation (which I built in 2011). That’s 12½ years and some extra hours uptime. They are all three fine and still in use.
That’s true. Nonetheless the maintainers seem to see AI as a valid way to reduce their workload.