I have a 4 node cluster. Each one has 1 SSD for the OS and block.dbs
(50Gb partition for each OSD), one 4Tb hdd and two 8Tb hdds.
I have 15% iowait average.
In any other server 15% seems too much. But ceph is a storage service
cluster.
There's a way to minimize the iowait or to better measure where is my
bottleneck? (¿Any of the HDDs maybe?)
Is time to stop changing my HDDs for bigger ones and add a new node?
--
Alfrenovsky
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
