i google provide guidelines on durability , availability guarantees provided cloud bigtable service. here understanding far: the fact minimum cluster requires 3 nodes suggests that, @ least within zone, data highly durable , replicated 3 nodes. however, this answer googler states "cloud bigtable doesn’t replicate data" — directly contradicting quote on cloud bigtable homepage claims "is built replicated storage strategy". it? replicated or not? , if so, how many copies kept? the fact clusters can set within particular zone suggests availability of cluster tied directly availability of zone. if want have highly available bigtable-based data storage, best practice set independent clusters across multiple zones , handle synchronisation of writes across clusters myself? there no information on whether bigtable clusters across zones independent or not. if set clusters across multiple zones, , 1 zone goes down, expect clusters in other zones carry on work...
Comments
Post a Comment