Open Stack

Hi,

I am trying to deploy cinder-backup service and c-vol uses multi-backend,
the configuration as follows:

...
enabled_backends = ceph1, ceph2
backup_driver = cinder.backup.drivers.ceph
backup_ceph_conf=/etc/ceph/ceph.conf
backup_ceph_user = admin
backup_ceph_chunk_size = 134217728
backup_ceph_pool = cinder-backups
backup_ceph_stripe_unit = 0
backup_ceph_stripe_count = 0
...
[ceph1]
rbd_ceph_conf=/etc/ceph/server-31.conf
backend_host=cinder
rbd_user=admin
volume_backend_name=ceph1
volume_driver=cinder.volume.drivers.rbd.RBDDriver
rbd_pool=cinder-volumes
rbd_secret_uuid=035adbab-a410-4ec6-a3f1-d1eaaac4db6a
rbd_store_chunk_size=4
rbd_cluster_name=server-31

[ceph2]
rbd_ceph_conf=/etc/ceph/server-32.conf
backend_host=cinder
rbd_user=admin
volume_backend_name=ceph2
volume_driver=cinder.volume.drivers.rbd.RBDDriver
rbd_pool=cinder-volumes
rbd_secret_uuid=7058cdfc-8297-4c06-b965-0604cc7991fc
rbd_store_chunk_size=4
rbd_cluster_name=server-32

the server-31 and server-32 are different ceph cluster.

All cinder services look OK:

cinder service-list
+------------------+-----------------+------+---------+-------+----------------------------+-----------------+
|      Binary      |       Host      | Zone |  Status | State |
Updated_at         | Disabled Reason |
+------------------+-----------------+------+---------+-------+----------------------------+-----------------+
|  cinder-backup   |    server-31    | nova | enabled |   up  |
2017-02-07T02:13:51.000000 |        -        |
| cinder-scheduler |    server-33    | nova | enabled |   up  |
2017-02-07T02:13:52.000000 |        -        |
|  cinder-volume   |   cinder at ceph1  | nova | enabled |   up  |
2017-02-07T02:13:43.000000 |        -        |
|  cinder-volume   |   cinder at ceph2  | nova | enabled |   up  |
2017-02-07T02:13:43.000000 |        -        |
+------------------+-----------------+------+---------+-------+----------------------------+-----------------+

Firstly, create two volume with different backend:

cinder create --volume-type ceph1 --name ceph1 1
cinder create --volume-type ceph2 --name ceph2 2

Then try to create backup:

cinder backup-create ceph1
cinder backup-create ceph2

The result show that the ceph1 is ok but the ceph2 always fails to create
backup, and the log show the image not found error:

2017-02-06 15:08:21.185 121910 ERROR os_brick.initiator.linuxrbd
[req-8abc2a96-1152-4906-8dfa-9bdb8f7001cd c12f90f5257a446682aa7cd11b2d1a97
78b0cf3fa582430ba9be966c7af441bc - - -] error opening rbd image
volume-380ba471-b7f0-4e8f-a55a-5e0c39fa1349
2017-02-06 15:08:21.185 121910 ERROR os_brick.initiator.linuxrbd Traceback
(most recent call last):
2017-02-06 15:08:21.185 121910 ERROR os_brick.initiator.linuxrbd   File
"/usr/lib/python2.7/site-packages/os_brick/initiator/linuxrbd.py", line 98,
in __init__
2017-02-06 15:08:21.185 121910 ERROR os_brick.initiator.linuxrbd
read_only=read_only)
2017-02-06 15:08:21.185 121910 ERROR os_brick.initiator.linuxrbd   File
"rbd.pyx", line 1042, in rbd.Image.__init__ (rbd.c:8581)
2017-02-06 15:08:21.185 121910 ERROR os_brick.initiator.linuxrbd
ImageNotFound: error opening image
volume-380ba471-b7f0-4e8f-a55a-5e0c39fa1349 at snapshot None
2017-02-06 15:08:21.185 121910 ERROR os_brick.initiator.linuxrbd

It seems that the backup service connected wrong ceph cluster to get volume
image.

I wonder if the backup service support multibackend or I did something
wrong configuration.

Thank you very much!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20170207/2db31d9a/attachment.html>