> 2. > If I remove the db device from the nvme with the ceph-objectstore-tool and keep it with the block, would it be an issue still? I guess if stay together cannot spillover anywhere. > I guess need to compact the spilledover disks before remove db. If you don't use a WAL/DB device, you are basically running with 100% spillover. In this case, the wal/db ends up on the data drive directly, instead of having some part of the wal/db on the (hopefully) faster device. > 3. > Correct me if I’m wrong but the separate db device is just a help for the block to be able to find the files from which block it is located in the disk so if I remove, the osd will still know where is the data, but with looking on the block device itself not on the separated nvme. No, while the DB part is holding vital data for the rocksdb/osd, if you skip using a separate one, that list is colocated with the data content on the main device. The idea is to have a faster device for those lists and write-buffers and all that, to allow more iops for small writes into said lists, and let the big IO operations go directly to the main device, allowing for more parallelism and a better experience overall. -- 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
 |