3 ms·
protect it... like running more than one instance of everybody so that we still experience continuity when a bolide strikes a compute node? Or like, we build a
by __MatrixMan__ 3y ago
protect it... like running more than one instance of everybody so that we still experience continuity when a bolide strikes a compute node? Or like, we build a wall?
- burnished 3y agoPor que no los dos?
- __MatrixMan__ 3y agoBecause it's unclear where you'd get the mass for the wall. Seems better to just let the nodes fail sometimes and tolerate the unreliability in a RAID sort of way.