Hi all, I'm trying to test pg_log mechanism under stress test by using a simple 3 nodes with 15 osds ceph cluster(replica size is 2). The following were test steps: 1. Set mon_osd_down_out_interval to 2 days 2. Using at least 10 threads FIO and librbd to send random 4KB r/w IOs continuously.
3. Shutdown one node, that means 5 osds were down, and pg_log mechanism started to work. After running 1.5 days, I observed two results: 1. There were no pg_log file on all running OSD disk 2. Almost every running OSD process occupied up to 1 GB memory. So my question was: It seemed like OSD kept all pg_log data in memory, was that true? If not, could someone help to explain what happen to this? BTW, XFS was used in this environment. Thanks, Jan
_______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
