On Mon, 23 Sep 2013, Dan van der Ster wrote: > The journals are on the same spinning disk & partition as the OSD > data. Each spinning disk is 1 OSD. Quick clarification: do you mean that the journal is a file on the same fs as osd_data? If so, moving it to a separate partition on the same disk usually helps quite a bit. The overhead of a write+fsync inside the fs tends to be much higher than an direct-io write to another partition. You might try measuring that directly and comparing it to the 33ms append+fsync that you previously saw. sage -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
