Den fre 10 sep. 2021 kl 12:56 skrev huxiaoyu@xxxxxxxxxxxx <huxiaoyu@xxxxxxxxxxxx>: > Dear Ceph folks, > This is closely related to my previous questions on how to do safely and reliabely RadosGW remote replication. > My major task is to backup S3 buckets. One obvious method is to use Ceph RadosGW multisite replication. I am wondering whether this is the best way to do S3 storage backup， or are there any better methods or alternatives? I am dealing with ca. 5-8TB amount of new data per day "best" is hard to answer when one doesn't know which of the many dimensions you are considering. There might be "fastest", "safest", "puts least load on secondary site", "transfers least amount of data per sync", "gives guarantees to always be consistent", "easiest to restart in the middle of a broken transfer", "is possible for non ceph-admin remote users to initiate" and probably ten more I can't imagine right now. Given that different data have different needs (an old index and new content might be ok, since all objects in the index are reachable, whereas newer index than data might be a disaster if you read the index before the data is actually in place) You will have to weigh several options against each other. Would you sync a few larger chunks (nights, weekends) or start resync as soon as one single object appears on the source? You need to know how to handle deletes, or find out when site A but not site B has data, or when only site B but not site A has. -- May the most significant bit of your life be positive. _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |