Ceph

While playing with the erasure code, I noticed that erasure coded objects were not cleaned up.
My setup:

• Debian Wheezy 7.4 with Kernel 3.10
• Ceph 0.80.1

How to reproduce:

$ sudo ceph osd pool create ec 1024 1024 erasure default
$ dd if=/dev/zero of=ec bs=100M count=1
$ sudo ceph -s
cluster 92775394-e114-40f5-8789-cd96220579f0
health HEALTH_OK
monmap e1: 1 mons at {ceph001=10.143.114.185:6789/0}, election epoch 2, quorum 0 ceph001
mdsmap e4: 1/1/1 up {0=ceph002=up:active}, 4 up:standby
osdmap e50: 29 osds: 29 up, 29 in
pgmap v118: 2240 pgs, 5 pools, 1884 bytes data, 20 objects
1765 MB used, 26771 GB / 26773 GB avail
2240 active+clean
$ sudo ceph df
GLOBAL:
SIZE AVAIL RAW USED %RAW USED
26773G 26771G 1765M 0
POOLS:
NAME ID USED %USED OBJECTS
data 0 0 0 0
metadata 1 1884 0 20
rbd 2 0 0 0
leseb 3 0 0 0
ec 4 0 0 0
$ rados -p ec put ec ec

$ sudo ceph -s
cluster 92775394-e114-40f5-8789-cd96220579f0
health HEALTH_OK
monmap e1: 1 mons at {ceph001=10.143.114.185:6789/0}, election epoch 2, quorum 0 ceph001
mdsmap e4: 1/1/1 up {0=ceph002=up:active}, 4 up:standby
osdmap e50: 29 osds: 29 up, 29 in
pgmap v119: 2240 pgs, 5 pools, 100 MB data, 21 objects
1796 MB used, 26771 GB / 26773 GB avail
2240 active+clean
client io 359 kB/s wr, 0 op/s
$ sudo ceph df
GLOBAL:
SIZE AVAIL RAW USED %RAW USED
26773G 26771G 1796M 0
POOLS:
NAME ID USED %USED OBJECTS
data 0 0 0 0
metadata 1 1884 0 20
rbd 2 0 0 0
leseb 3 0 0 0
ec 4 102400k 0 1

Everything is fine.
Now deleting the object:

$ rados -p ec rm ec

$ sudo ceph -s
cluster 92775394-e114-40f5-8789-cd96220579f0
health HEALTH_OK
monmap e1: 1 mons at {ceph001=10.143.114.185:6789/0}, election epoch 2, quorum 0 ceph001
mdsmap e4: 1/1/1 up {0=ceph002=up:active}, 4 up:standby
osdmap e50: 29 osds: 29 up, 29 in
pgmap v126: 2240 pgs, 5 pools, 1884 bytes data, 20 objects
2322 MB used, 26770 GB / 26773 GB avail
2240 active+clean
client io 0 B/s wr, 0 op/s

$ sudo ceph df
GLOBAL:
SIZE AVAIL RAW USED %RAW USED
26773G 26770G 2322M 0
POOLS:
NAME ID USED %USED OBJECTS
data 0 0 0 0
metadata 1 1884 0 20
rbd 2 0 0 0
leseb 3 0 0 0
ec 4 0 0 0

Everything looks good, however:

$ sudo ceph osd map ec ec
osdmap e50 pool 'ec' (4) object 'ec' -> pg 4.f79165da (4.1da) -> up ([28,26,11], p28) acting ([28,26,11], p28)

root@ceph003:/var/lib/ceph/osd/ceph-11/current/4.1das2_head# ls alh
total 51M
drwxr-xr-x 2 root root 38 May 26 16:48 .
drwxr-xr-x 245 root root 8.0K May 26 16:04 ..
-rw-r--r- 1 root root 50M May 26 16:46 ec__head_F79165DA__4_80_2

root@ceph006:/var/lib/ceph/osd/ceph-28/current/4.1das0_head# ls lah
total 51M
drwxr-xr-x 2 root root 38 May 26 16:48 .
drwxr-xr-x 248 root root 8.0K May 26 16:04 ..
-rw-r--r- 1 root root 50M May 26 16:46 ec__head_F79165DA__4_80_0

root@ceph005:/var/lib/ceph/osd/ceph-26/current/4.1das1_head# ls alhtotal 51M
drwxr-xr-x 2 root root 38 May 26 16:48 .
drwxr-xr-x 231 root root 8.0K May 26 16:04 ..
-rw-r--r- 1 root root 50M May 26 16:46 ec__head_F79165DA__4_80_1

Am I missing something?
Btw I tried with a replicated pool and I don't have this problem.

Cheers.