Hi, You might find this blog post helpful: IBM Storage Ceph Object Storage Multisite - Part 1 <https://community.ibm.com/community/user/storage/blogs/daniel-alexander-parkes/2024/01/23/ibm-storage-ceph-object-storage-multisite-part1> . The number of RGWs and their configurations depends on the load, the data being replicated across zones, and the resources available on the RGW servers, such as network capacity. In my setup, I have four zones replicated with two dedicated daemons per RGW and three RGWs per cluster. On Wed, Dec 18, 2024 at 9:47 AM Adam Prycki <aprycki@man.poznan.pl> wrote:
Hello,
I was recently reading rook ceph multisite documentation and I've found this rgw recommendation:
"Scaling the number of gateways that run the synchronization thread to 2 or more can increase the latency of the replication of each S3 object. The recommended way to scale a multisite configuration is to dissociate the gateway dedicated to the synchronization from gateways that serve clients."
Are there any official ceph recommendations about how many RGWs with rgw_run_sync_thread=true we should deploy in multisite zone?
Have you ever experienced issues related to this in production?
My plan for our multisite was to create 3 RGW daemons with rgw_run_sync_thread, put them behind load balancer and put load balancer VIP in zone endpoint. Rook documentation suggests it would be sub optimal configuration.
Best regards Adam Prycki _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io