Ceph » CephFS

Reproducing this in a vstart environment:

1. Mount a client
2. in python, do "f = open('mnt/foo.bin', 'w')"
3. Leave the python shell open and do a umount -f on the mount
4. The ceph-fuse process segfaults.

 ceph version v11.0.0-171-gc71cb0a (c71cb0adadaf22dd8a2c8e429afc6d2e05b8eda9)
 1: (ceph::BackTrace::BackTrace(int)+0x2d) [0x11393a1]
 2: bin/ceph-fuse() [0x1138a5a]
 3: (()+0x10430) [0x7f9237e2f430]
 4: (Inode::caps_issued(int*) const+0x23) [0x112c6f7]
 5: (operator<<(std::ostream&, Inode const&)+0x39) [0x112b709]
 6: (Client::_release_fh(Fh*)+0x2c3) [0x105a965]
 7: (Client::unmount()+0x7c7) [0x1046b31]
 8: (main()+0xbe1) [0x10072c6]
 9: (__libc_start_main()+0xf0) [0x7f92359ba700]
 10: (_start()+0x29) [0x1005d49]
 NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.

Note that I'm using --client-die-on-failed-remount=false, I can't immediately see how that would be tied to this but it's perhaps possible.