Hi all,
Regarding this problem, I tried something that works for me:
I noticed that using Contrail v2.10 I cannot install the cinder-volume
service because this package is not available, so I decided to create my
own package.
These are the steps I followed in Ubuntu 14.04.1:
I'm using LVM with a volume group named cinder-volumes to store the volumes.
# 1. Create my cinder-volume_2014.1.3-0ubuntu1~cloud0_all.deb package
# 1.1. download original cinder package
https://launchpadlibrarian.net/187842740/cinder-volume_2014.1.3-0ubuntu1.1_all.deb
# 1.2. create temporal folder
mkdir mycinder
cd mycinder
# 1.3. extract package files
dpkg-deb -x ../cinder-volume_2014.1.3-0ubuntu1.1_all.deb .
# 1.4. extract package control files
dpkg-deb -e ../cinder-volume_2014.1.3-0ubuntu1.1_all.deb
# 1.5. edit file DEBIAN/control and change package version and
dependency in cinder-common
# replace "2014.1.3-0ubuntu1.1" by "1:2014.1.3-0ubuntu1~cloud0" so next
lines should result
Version: 1:2014.1.3-0ubuntu1~cloud0
Depends: cinder-common (= 1:2014.1.3-0ubuntu1~cloud0), lvm2, tgt,
sysv-rc (>= 2.88dsf-24) | file-rc (>= 0.8.16), python:any
Keep the rest of file as is
# 1.6. generate new package
dpkg-deb -b . ../cinder-volume_2014.1.3-0ubuntu1~cloud0_all.deb
dpkg-deb -i . ../cinder-volume_2014.1.3-0ubuntu1~cloud0_all.deb
# 2. Configure database in openstack node where cinder api/scheduler is
installed:
# 2.1. Take root password used for the database creation
grep "mysql -u root --password=" /opt/contrail/utils/setup_*
# 2.2. Login into database and grant permissions to cinder user
mysql -u root --password=<password taken from previous section>
mysql> GRANT ALL PRIVILEGES ON cinder.* TO 'cinder'@'localhost'
IDENTIFIED BY 'cinder';
mysql> GRANT ALL PRIVILEGES ON cinder.* TO 'cinder'@'<your openstack
node hostname>' IDENTIFIED BY 'cinder';
mysql> GRANT ALL PRIVILEGES ON cinder.* TO 'cinder'@'%' IDENTIFIED BY
'cinder';
mysql> exit
# 2.3. Populate cinder database
su -s /bin/sh -c "cinder-manage db sync" cinder
# 3. Install package
# 3.1. Copy new package cinder-volume_2014.1.3-0ubuntu1~cloud0_all.deb
into compute nodes
# 3.2. Configure cinder api/scheduler
nano /etc/cinder/cinder.conf
[DEFAULT]
rootwrap_config = /etc/cinder/rootwrap.conf
api_paste_config = /etc/cinder/api-paste.ini
iscsi_helper = tgtadm
volume_name_template = volume-%s
volume_group = cinder-volumes
verbose = True
auth_strategy = keystone
state_path = /var/lib/cinder
lock_path = /var/lock/cinder
volumes_dir = /var/lib/cinder/volumes
rpc_backend = rabbit
rabbit_host = <your_openstack_node_IP>
rabbit_port = 5672
my_ip = <your_openstack_node_IP>
glance_host = <your_openstack_node_IP>
osapi_volume_workers = 4
[database]
connection =
mysql://cinder:cinder@<your_cinder_api/scheduler_node_IP>/cinder
[keystone_authtoken]
admin_tenant_name = service
admin_user = cinder
admin_password = (copy the value from same attribute in
/etc/nova/nova.conf)
auth_protocol = http
auth_host = <your_openstack_node_IP>
auth_port = 35357
auth_uri = http://<your_openstack_node_IP>:5000
# 3.3. Restart cinder-api and cinder-scheduler services in openstack node
service cinder-api restart
service cinder-scheduler restart
# 4. Install new cinder-volume in compute nodes
# 4.1. copy cinder-volumes package into contrail install repo
cp cinder-volume_2014.1.3-0ubuntu1~cloud0_all.deb
/opt/contrail/contrail_install_repo
# 4.2. move to contrail install repo folder
cd /opt/contrail/contrail_install_repo
# 4.3. regenerate Packages.gz
dpkg-scanpackages . /dev/null | gzip -9c > Packages.gz
# 4.4. update apt-get cache and install cinder-volume
apt-get update
apt-get install cinder-volume
# 5. Configure compute node
# 5.1. Configure cinder-volume service
nano /etc/cinder/cinder.conf
[DEFAULT]
rootwrap_config = /etc/cinder/rootwrap.conf
api_paste_config = /etc/cinder/api-paste.ini
iscsi_helper = tgtadm
volume_name_template = volume-%s
volume_group = cinder-volumes
#verbose = True
auth_strategy = keystone
state_path = /var/lib/cinder
lock_path = /var/lock/cinder
volumes_dir = /var/lib/cinder/volumes
rpc_backend = rabbit
rabbit_host = <your_openstack_node_IP>
rabbit_port = 5672
my_ip = <your_compute_node_IP>
glance_host = <your_openstack_node_IP>
[database]
connection =
mysql://cinder:cinder@<your_openstack_node_IP>/cinder
[keystone_authtoken]
admin_tenant_name = service
admin_user = cinder
admin_password = (copy the value from same attribute in
/etc/nova/nova.conf)
auth_protocol = http
auth_host = <your_openstack_node_IP>
auth_port = 35357
auth_uri = http://<your_openstack_node_IP>:5000
# 5.2. Configure LVM service
nano /etc/lvm/lvm.conf
Leave file as is, except for the filter attribute in
devices section.
The device specified (sda4) has to be the same where
cinder-volumes resides.
devices {
...
# Cinder filter
filter = [ "a/sda4/", "r/.*/" ]
...
}
# 5.3. Restart cinder-volume and tgt services
service cinder-volume restart
service tgt restart
Right now you should be able to create instances specifying to create a
volume from an image or create empty volumes without problem.
The only issue I'm experiencing is that Openstack's Dashboard, in Admin
> Hypervisors reports an incorrect storage total space and used space.
Except for this, it works correctly.
Hope it helps.
Lluis
El 14/04/15 a les 18:04, Joao Prino ha escrit:
Hi Saravanan,
Thanks for the update.
The output:
# dpkg -l| grep cinder
ii cinder-api 1:2014.1.3-0ubuntu1~cloud0
all Cinder storage service - API server
ii cinder-common 1:2014.1.3-0ubuntu1~cloud0
all Cinder storage service - common files
ii cinder-scheduler 1:2014.1.3-0ubuntu1~cloud0
all Cinder storage service - Scheduler server
ii python-cinder 1:2014.1.3-0ubuntu1~cloud0
all Cinder Python libraries
ii python-cinderclient 1:1.0.8-0ubuntu1
all python bindings to the OpenStack Volume API
That might explain the behaviour, let me install the packages and will
get back to you.
Cheers,
Joao
-----Original Message-----
From: Saravanan Purushothaman [mailto:[email protected]]
Sent: 14 April 2015 16:56
To: Joao Prino; Lluís Gifre; ask-contrail; [email protected]
Cc: contrailuser
Subject: RE: [Users] Error creating volumes with cinder
Hi Joao,
Can you try this.. "dpkg -l| grep cinder"
I think cinder-volume package is not installed.
if cinder-volume service is not running then we may get this "No
valid host was found"
NOTE:
cinder-volume is storage related package. If you
want storage on contrail then please follow the link instructions to install
contrail-storage package. (cinder-volume will be installed as part of
contrail-storage package)
you can get the "contrail storage packages" from this link http://www.juniper.net/support/downloads/?p=contrail#sw
Regards,
Saravanan
-----Original Message-----
From: Joao Prino [mailto:[email protected]]
Sent: Tuesday, April 14, 2015 12:37 AM
To: Lluís Gifre; ask-contrail; [email protected]
Cc: contrailuser
Subject: RE: [Users] Error creating volumes with cinder
Hello Lluis Dev,
I'm calling Dev's attention to this problem.
Look forward to hear back whenever possible.
Cheers,
Joao
-----Original Message-----
From: Lluís Gifre [mailto:[email protected]]
Sent: 10 April 2015 17:03
To: Joao Prino; Adrian Smith
Cc: contrailuser
Subject: Re: [Users] Error creating volumes with cinder
Hi Joao,
Yes, I have already done this correciton but the problem "No backend provided" and
"No valid host was found" while the cinder-scheduler is creating the volume is still
happening.
Thanks,
Lluis
El 10/04/15 a les 17:54, Joao Prino ha escrit:
Hi Lluis,
I had this exact same problem where was able to solve it by adding the "rabbit_host =
<ip>" under '/etc/cinder/cinder.conf', editing '/etc/rabbitmq/rabbitmq.config' to match my cluster
setup's fqdn (wrongly populated) and changed the existing "rabbit_host = <ip>" from
'/etc/nova/nova.conf as it had the wrong IP assigned to it.
This problem was due to having two NICs where RabbitMQ was given my
NIC1 address settings and it should have had my NIC2's address according to the
testbed file..
Hope it helps!
Joao
-----Original Message-----
From: Users [mailto:[email protected]] On Behalf Of
Lluís Gifre
Sent: 10 April 2015 15:19
To: Adrian Smith
Cc: contrailuser
Subject: Re: [Users] Error creating volumes with cinder
Hi Adrian,
Thanks for your quick answer.
I checked the log you mentioned but I hadn't found any interesting thing.
Do you have any other idea?
Thanks,
Lluis
El 10/04/15 a les 13:14, Adrian Smith ha escrit:
Hi Lluis,
Take a look in the cinder scheduler log,
/var/log/cinder/scheduler.log. It should have a more meaningful error.
Adrian
On 10 April 2015 at 11:48, Lluís Gifre <[email protected]> wrote:
Dear all,
I'm deploying a opencontrail 2.10 + openstack icehouse testbed in
Ubuntu 14.04.
I tried it in many ways (with physical and virtual machines, single
and multi-box testbeds).
Right now I'm trying with a single virtual machine with LVM and the
cinder-volumes volume group created.
The deploy completes successfully.
However, just after completing the set-up, I realized that in
cinder-schedule.log the next message is added periodically:
2015-04-10 07:27:29.426 2315 INFO
oslo.messaging._drivers.impl_rabbit
[req-e77c84ab-49f6-40f2-bb20-1c62b23605d6 - - - - -] Reconnecting to
AMQP server on localhost:5672
2015-04-10 07:27:29.427 2315 INFO
oslo.messaging._drivers.impl_rabbit
[req-e77c84ab-49f6-40f2-bb20-1c62b23605d6 - - - - -] Delaying
reconnect for
1.0 seconds...
2015-04-10 07:27:30.442 2315 ERROR
oslo.messaging._drivers.impl_rabbit
[req-e77c84ab-49f6-40f2-bb20-1c62b23605d6 - - - - -] AMQP server on
localhost:5672 is unreachable: [Errno 111] ECONNREFUSED. Trying
again in 30 seconds.
To solve the problem I edited /etc/cinder/cinder.conf and added:
rabbit_host = 192.168.67.13 # IP address taken from
/etc/rabbitmq/rabbitmq.config rabbit_port = 5672
Content of my /etc/rabbitmq/rabbitmq.config:
[
{rabbit, [ {tcp_listeners, [{"192.168.67.13", 5672}]},
{loopback_users, []},
{log_levels,[{connection, info},{mirroring, info}]} ]
}
].
After rebooting, this change solved the recurrent reconnecting to
AMQP problem.
Right now the cinder-scheduler log shows:
2015-04-10 07:32:44.063 2525 AUDIT cinder.service [-] Starting
cinder-scheduler node (version 2014.1.3)
2015-04-10 07:32:44.197 2525 INFO
oslo.messaging._drivers.impl_rabbit
[req-21e86bf9-3fa6-4ab0-9968-616242200c41 - - - - -] Connected to
AMQP server on 192.168.67.13:5672
2015-04-10 07:32:48.383 2525 INFO
oslo.messaging._drivers.impl_rabbit
[-] Connected to AMQP server on 192.168.67.13:5672
Next I tried to create a volume, no mater if I do it from the
openstack dashboard or using cinder's or nova's command line
interface, the result is an error. For example trying with cinder command line:
# cinder list
+----+--------+--------------+------+-------------+----------+-------------+
| ID | Status | Display Name | Size | Volume Type | Bootable |
| Attached to |
+----+--------+--------------+------+-------------+----------+-------------+
+----+--------+--------------+------+-------------+----------+-------------+
# cinder create --display-name test 1
+---------------------+--------------------------------------+
| Property | Value |
+---------------------+--------------------------------------+
| attachments | [] |
| availability_zone | nova |
| bootable | false |
| created_at | 2015-04-10T07:57:12.787493 |
| display_description | None |
| display_name | test |
| encrypted | False |
| id | d374b571-5df1-47f3-ae6e-c3218aebb9db |
| metadata | {} |
| size | 1 |
| snapshot_id | None |
| source_volid | None |
| status | creating |
| volume_type | None |
+---------------------+--------------------------------------+
# cinder list
+--------------------------------------+--------+--------------+------+-------------+----------+-------------+
| ID | Status | Display Name | Size |
Volume Type | Bootable | Attached to |
+--------------------------------------+--------+--------------+------+-------------+----------+-------------+
| d374b571-5df1-47f3-ae6e-c3218aebb9db | error | test | 1 |
None | false | |
+--------------------------------------+--------+--------------+------+-------------+----------+-------------+
Then I checked the /var/log/cinder/cinder-scheduler.log and found
next
message:
2015-04-10 07:36:14.547 2525 WARNING cinder.context [-] Arguments
dropped when creating context: {'user':
u'ad2be0cd57c44177b1ba9b0735ea0f44',
'tenant': u'90aab64e7e2b45f3891b0f08978e063e', 'user_identity':
u'ad2be0cd57c44177b1ba9b0735ea0f44 90aab64e7e2b45f3891b0f08978e063e
-
- -'}
2015-04-10 07:36:14.629 2525 ERROR
cinder.scheduler.flows.create_volume
[req-ba5d16aa-5092-4f5f-a4c6-d9481df83bfd
ad2be0cd57c44177b1ba9b0735ea0f44 90aab64e7e2b45f3891b0f08978e063e -
- -] Failed to schedule_create_volume: No valid host was found.
I have enabled DEBUG by adding debug = True to
/etc/cinder/cinder.conf and found in the cinder-scheduler.log: ... No backend
provided ...
I installed and condifured tgtd according to
http://rconradharris.com/2013/01/14/getting-cinder-up-and-running.ht
m
l
But the error still remains.
Do somebody have an idea of how to solve this problem?
Thanks,
Lluis
_______________________________________________
Users mailing list
[email protected]
http://lists.opencontrail.org/mailman/listinfo/users_lists.opencontr
a
il.org
_______________________________________________
Users mailing list
[email protected]
http://lists.opencontrail.org/mailman/listinfo/users_lists.opencontrai
l.org
_______________________________________________
Users mailing list
[email protected]
http://lists.opencontrail.org/mailman/listinfo/users_lists.opencontrail.org
