Hi there,
We are looking at using Ceph (Jewel) for a use case that is very write heavy strictly as an object store. We've been working with Rados Gateway because we can easily integrate with existing S3 libraries... but we will never be doing any of the bucket listing operations. I am concerned about the potential bottleneck of the RGW index files.
I've read here that Jewel now supports "Blind Buckets"... with some reference to setting a the RGWBucketIndexType to RGWBIType_Indexless... and I'm guessing its set as "index_type" here. In the docs, the only "index_type" reference I see is here, under the placement pools. However, the Pools documentation doesn't really give a clue how to set this value, or even if this is the proper index_type field that I'm guessing.
So the two things I'm interested in figuring out is:
1) Are "Blind Buckets" actually production ready
2) How do I configure Rados Gateway to omit the index files?
--Tyler
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
