http://bugzilla.suse.com/show_bug.cgi?id=1002455 http://bugzilla.suse.com/show_bug.cgi?id=1002455#c1 Nathan Cutler <ncutler@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ddiss@suse.com --- Comment #1 from Nathan Cutler <ncutler@suse.com> ---

0> 2016-10-02 13:07:13.496806 7fdec3619700 -1 *** Caught signal (Segmentation fault) ** in thread 7fdec3619700 thread_name:ceph-osd

ceph version 10.2.2-368-g6bc382e (6bc382e6ce0bad619ae200604a0d27451a275ba9) 1: (()+0x901d22) [0x55a562d32d22] 2: (()+0x10b00) [0x7fdec8027b00] 3: [0x55a56d9f41b0] NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.

Please install ceph-osd-debuginfo and reproduce again. In general, the ceph-osd process should never crash, so this is a bug. But we will need to get some more information before we can try to reproduce it. Can you provide a description of your cluster? Is the cluster running in HEALTH_OK before you attempt this? Are you stopping the OSD ("systemctl stop ceph-osd@1.service") before you run ceph-osd --flush-journal on it? Can you provide dumps of the partition tables of the OSD data and journal disks using "sgdisk -p" ? In addition, can you post the output of "sgdisk --info=$PARTITION_NUMBER $DEVICE" for all the partitions on the OSD disk and the external journal disk? (So we can see the partition GUID codes.) Can you give us access to the cluster if need be? David, any other info we need? -- You are receiving this mail because: You are on the CC list for the bug.