Ceph

This happened after moving a journal from a file to an ssd, and changing filestore xattr use omap from true to false, with the wip-f branch.
4 osds crashed with this backtrace. A log with osd and filestore debugging is attached.
There's a core file on metropolis:~joshd/core-bad-object-info

 ceph version 0.48.2argonaut-16-g24fc459 (commit:24fc4599a7d1d3d49ce4623723723bd31f701cca)
 1: /usr/bin/ceph-osd() [0x6f9f0a]
 2: (()+0xfcb0) [0x7ffbd7ebbcb0]
 3: (gsignal()+0x35) [0x7ffbd63a8425]
 4: (abort()+0x17b) [0x7ffbd63abb8b]
 5: (__gnu_cxx::__verbose_terminate_handler()+0x11d) [0x7ffbd6cfa69d]
 6: (()+0xb5846) [0x7ffbd6cf8846]
 7: (()+0xb5873) [0x7ffbd6cf8873]
 8: (()+0xb596e) [0x7ffbd6cf896e]
 9: (ceph::buffer::list::iterator::copy(unsigned int, char*)+0x127) [0x7b6ca7]
 10: (object_info_t::decode(ceph::buffer::list::iterator&)+0x2f) [0x81793f]
 11: (ReplicatedPG::send_push(int, ObjectRecoveryInfo, ObjectRecoveryProgress, ObjectRecoveryProgress*)+0xe47) [0x54d4d7]
 12: (ReplicatedPG::sub_op_pull(std::tr1::shared_ptr<OpRequest>)+0x342) [0x5508a2]
 13: (ReplicatedPG::do_sub_op(std::tr1::shared_ptr<OpRequest>)+0x347) [0x559087]
 14: (PG::do_request(std::tr1::shared_ptr<OpRequest>)+0x9f) [0x60d9ef]
 15: (OSD::dequeue_op(PG*)+0x238) [0x5cca38]
 16: (ThreadPool::worker()+0x4d5) [0x7acfe5]
 17: (ThreadPool::WorkThread::entry()+0xd) [0x5e584d]
 18: (()+0x7e9a) [0x7ffbd7eb3e9a]
 19: (clone()+0x6d) [0x7ffbd6465cbd]