3 ms·
Nice post and inspiring, but one small point bother me: "I feel GREAT not living in FEAR". Well, you just traded the fear of being shut down by big corpo for th
by julienfr112 6y ago
Nice post and inspiring, but one small point bother me: "I feel GREAT not living in FEAR". Well, you just traded the fear of being shut down by big corpo for the fear of having your hard disk or memory or fan or whatever fail.
- nacs 6y agoMost rackmount servers double up on everything like power-supplies. Harddrives should be on RAID so a single drive failure shouldn't be a problem and you would have some time to goto the colo and switch out the hard drive. CPU and RAM last a very long time so it shouldn't be a problem (I've had literally 1 RAM stick failure in over a decade of hosting and the datacenter swapped out the bad memory stick in less than 30 minutes -- it also didn't bring down the server either, the kernel log started showing a bunch of ECC issues and a quick memory check pointed to the stick that had the problem). And of course you can just get double or triple the amount of servers with IP failover and such to be even more resilient to hardware failures. This article is just showing how to get started. Throw more servers at it for more redundancy and scaling as needed.