On 02/09/13 07:19, Fuchs, Andreas (SwissTXT) wrote:
Reading through the documentation and talking to several peaople leads to the conclusion that it's a best practice to place the journal of an OSD instance to a separate SSD disk to speed writing up. But is this true? i have 3 new dell servers for testing available with 12 x 4 TB SATA and 2 x 100GB SSD disks. I don't have the exact specs at hand but tests show: The SATA's sequential write speed is 300MB/s The SSD which is in RAID1 config is only 270MB/s ! was probably not the most expensive. When we put the journals on the OSD's i can expect a sequential wtite speed of 12 x 150MB/s (on write to journal, one to disk), this is 1800MB/s per Node. So to speed up the disks we need at least SSD's with a total seq write speed of 1800MB/s and to get the max out of the SATA's we even need 3600MB/s, this will become expensive! We are not shure how random the whole writing will get with real usage and how we could add this to the calculations. But one cheap SSD with a couple of journals for shure kills the write performance. How do you calculate the SSD sizing?
The thing to grasp is that SSD random and (f)sync performance is typically vastly superior to spinning disk. So while corresponding purely sequential performance may be similar (or even inferior) from an SSD - using one for a journal is usually a considerable advantage.
Regards Mark _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
