settingsLogin | Registersettings

[Openstack] Create instance fails on creating block device - Block Device Mapping is Invalid

0 votes

I'm trying my newly installed Openstack system and I'm getting
problem in starting my first instance.

----- s n i p -----
Build of instance 5193c2d9-0aaf-4f84-b108-f6884d97b571 aborted: Block Device Mapping is Invalid.
File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 1926, in dobuildandruninstance filterproperties) File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 2083, in buildandruninstance 'create.error', fault=e) File "/usr/lib/python2.7/dist-packages/osloutils/excutils.py", line 220, in exit self.forcereraise() File "/usr/lib/python2.7/dist-packages/osloutils/excutils.py", line 196, in forcereraise six.reraise(self.type_, self.value, self.tb) File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 2048, in buildandruninstance blockdevicemapping) as resources: File "/usr/lib/python2.7/contextlib.py", line 17, in enter return self.gen.next() File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 2206, in buildresources reason=e.format_message())
----- s n i p -----

Cleaning up all the logs for irrelevant stuff, I see:

----- s n i p -----
INFO cinder.api.v2.volumes Create volume of 5 GB
INFO cinder.volume.api Volume created successfully.
INFO cinder.volume.flows.manager.createvolume Volume 6b1dace4-78e1-452b-a455-c0fc882374f3: being created as image with specification: {'status': u'creating', 'imagelocation': (None, None), 'volumesize': 5, 'volumename': 'volume-6b1dace4-78e1-452b-a455-c0fc882374f3', 'imageid': u'8c15b5e8-9a67-4784-ad7a-0b1cc7b0bdec', 'imageservice': <cinder.image.glance.GlanceImageService object at 0x7fa4f31d8ad0>, 'imagemeta': {'status': u'active', 'name': u'fedora23', 'deleted': False, 'containerformat': u'docker', 'createdat': datetime.datetime(2016, 6, 15, 20, 38, 43, tzinfo=<iso8601.Utc>), 'diskformat': u'qcow2', 'updatedat': datetime.datetime(2016, 6, 15, 20, 38, 45, tzinfo=<iso8601.Utc>), 'id': u'8c15b5e8-9a67-4784-ad7a-0b1cc7b0bdec', 'owner': u'd524c8dfd9e9449798ebac9b025f8de6', 'minram': 0, 'checksum': u'38d62e2e1909c89f72ba4d5f5c0005d5', 'mindisk': 0, 'ispublic': True, 'deletedat': None, 'properties': {u'hypervisortype': u'docker', u'architecture': u'x8664'}, 'size': 234363392}}
INFO cinder.image.image
utils Image download 223.00 MB at 35.35 MB/s
WARN manila.context [-] Arguments dropped when creating context: {u'readonly': False, u'domain': None, u'showdeleted': False, u'useridentity': u'- - - - -', u'projectdomain': None, u'resourceuuid': None, u'userdomain': None}.
WARN manila.context [-] Arguments dropped when creating context: {u'readonly': False, u'domain': None, u'showdeleted': False, u'useridentity': u'- - - - -', u'projectdomain': None, u'resourceuuid': None, u'userdomain': None}.
INFO cinder.image.imageutils Converted 3072.00 MB image at 31.59 MB/s
INFO cinder.volume.flows.manager.create
volume Volume volume-6b1dace4-78e1-452b-a455-c0fc882374f3 (6b1dace4-78e1-452b-a455-c0fc882374f3): created successfully
INFO cinder.volume.manager Created volume successfully.
INFO cinder.api.v2.volumes Delete volume with id: 6b1dace4-78e1-452b-a455-c0fc882374f3
INFO cinder.volume.api Delete volume request issued successfully.
INFO eventlet.wsgi.server 10.0.4.5 "DELETE /v2/d524c8dfd9e9449798ebac9b025f8de6/volumes/6b1dace4-78e1-452b-a455-c0fc882374f3 HTTP/1.1" status: 202 len: 211 time: 0.1300900
INFO cinder.volume.targets.iscsi Skipping removeexport. No iscsitarget is presently exported for volume: 6b1dace4-78e1-452b-a455-c0fc882374f3
INFO cinder.volume.utils Performing secure delete on volume: /dev/mapper/blade_center-volume--6b1dace4--78e1--452b--a455--c0fc882374f3
----- s n i p -----

Full log at http://bayour.com/misc/openstack_instance_create-log.txt.

The web GUI say (this might be from another test, but I always
get the same):

----- s n i p -----
Error: Failed to perform requested operation on instance
"jessie-test", the instance has an error status: Please try again
later [Error: Build of instance a4e1deaa-cdf0-4fc7-8c54-579868c962c3
aborted: Block Device Mapping is Invalid.].
----- s n i p -----

I can see nothing relevant in this that would make it fail!
The only thing that bought my eye was that it isn't removing
the iSCSI target, because there isn't one..

This is (most of) my cinder.conf file:

----- s n i p -----
[DEFAULT]
myip = 10.0.4.1
storage
availabilityzone = nova
default
availabilityzone = nova
enabled
backends = lvm
iscsitargetprefix = iqn.2010-10.org.openstack:
iscsiipaddress = $myip
iscsi
port = 3260
iscsiiotype = blockio
iscsi
writecache = on
volume
group = bladecenter
scheduler
driver = cinder.scheduler.filter_scheduler.FilterScheduler

[lvm]
volumedriver = cinder.volume.drivers.lvm.LVMVolumeDriver
volume
group = bladecenter
iscsi
protocol = iscsi
iscsi_helper = tgtadm
----- s n i p -----

PS. Creating the instance from an already existing, empty
volume, didn't work either. Same message, and even less
information in the log.
--
As soon as you find a product that you really like,
they will stop making it.
- Wilson's Law


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
asked Jun 16, 2016 in openstack by Turbo_Fredriksson (8,980 points)   7 13 19

34 Responses

0 votes

Hi Turbo,

The first thing I want to know

  • Which VM are you creating(i.e; which OS image are you taking)?
  • What size are you using and all?

Secondly,

  • Which flavor are you using for VM creation.?

On Tue, Jun 21, 2016 at 12:56 PM, Eugen Block eblock@nde.ag wrote:

Can't you boot an instance without cinder?

>

Don't know, can I??

Well, you should ;-) How do you try to boot your instance, from CLI or
Horizon? If it's Horizon, you would have to NOT klick the button "Create a
new volume --> Yes" ;-) If it's CLI it's sufficient to only execute "nova
boot --flavor --image --nic net-id= (optional:
only if you have multiple networks available) "
This way you avoid creating a volume.

You could edit nova.conf

>
How?

It's usually the default, although I'm really not an expert in Openstack.
But if you simply try to set up nova on control and compute node following
an install guide, it should bring you there.
I followed
http://docs.openstack.org/mitaka/install-guide-obs/nova-controller-install.html,
there aren't many options to configure and it defaults to local file
storage.

From what I can see, it doesn't even start sharing via iSCSI
>

You should try to fix that before you try to use it with openstack.

Didn't even knew you could do that. Thought you HAD to use cinder/swift..

Please point me to a faq/howto/doc on how to do that, thanx!

I used this guide:

http://docs.openstack.org/mitaka/install-guide-obs/environment-networking-storage-cinder.html
In the section for block storage it says "Block storage node (Optional)",
so you wouldn't have to, but I guess it makes sense in the longterm. But as
I already said, first you should try to get an instance running at all
before using another backend.

Regards,
Eugen

Zitat von Turbo Fredriksson turbo@bayour.com:

On Jun 20, 2016, at 3:27 PM, Eugen Block wrote:

Can't you boot an instance without cinder?
>

Don't know, can I??

You could edit nova.conf to use local file system, just to have a running

instance. If that works you can switch to another backend.

How?

cinder create --image --name
>

I'll try that thanx. How do you do that with the "openstack" command?

Try debugging your iscsi connection, maybe first without openstack.
>

From what I can see, it doesn't even start sharing via iSCSI..

In my environment, I first tried to get all services running and working

without external backends, cinder, glance and nova all ran on local storage.

Didn't even knew you could do that. Thought you HAD to use cinder/swift..

Please point me to a faq/howto/doc on how to do that, thanx!

Then I tried other backends for cinder (iscsi), now all services use ceph.
>

ceph?
--
Life sucks and then you die


Mailing list:
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe :
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

--
Eugen Block voice : +49-40-559 51 75
NDE Netzdesign und -entwicklung AG fax : +49-40-559 51 77
Postfach 61 03 15
D-22423 Hamburg e-mail : eblock@nde.ag

    Vorsitzende des Aufsichtsrates: Angelika Mozdzen
      Sitz und Registergericht: Hamburg, HRB 90934
              Vorstand: Jens-U. Mozdzen
               USt-IdNr. DE 814 013 983


Mailing list:
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe :
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

--

Thanks & Regards,
Abhishek
*Cloudbyte Inc. *


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
responded Jun 21, 2016 by Abhishek_Shrivastava (3,180 points)   2 7
0 votes

On Jun 21, 2016, at 11:40 AM, Abhishek Shrivastava wrote:

The first thing I want to know

  • Which VM are you creating(i.e; which OS image are you taking)?

I've tried both the CirrOS and Debian GNU/Linux Jessie images.

http://download.cirros-cloud.net/0.3.4/cirros-0.3.4-x86_64-disk.img
http://cdimage.debian.org/cdimage/openstack/8.5.0/debian-8.5.0-openstack-amd64.qcow2

  • What size are you using and all?

Size? I've tried creating a volume from those images from 2GB to 20GB.

  • Which flavor are you using for VM creation.?

My own take on the m1.flavor:

openstack flavor create --ram 1024 --disk 10 --vcpus 1 --disk 5 m1.tiny
--
I love deadlines. I love the whooshing noise they
make as they go by.
- Douglas Adams

responded Jun 21, 2016 by Turbo_Fredriksson (8,980 points)   7 13 19
0 votes

​Have you tried any other flavors?

For instance if you are creating a 1GB volume then you can go for flavor
m1.tiny flavor.

So try creating a VM having boot volume size 1GB​ and use flavor m1.tiny
and see if it works.

On Tue, Jun 21, 2016 at 4:36 PM, Turbo Fredriksson turbo@bayour.com wrote:

On Jun 21, 2016, at 11:40 AM, Abhishek Shrivastava wrote:

The first thing I want to know

  • Which VM are you creating(i.e; which OS image are you taking)?

I've tried both the CirrOS and Debian GNU/Linux Jessie images.

http://download.cirros-cloud.net/0.3.4/cirros-0.3.4-x86_64-disk.img

http://cdimage.debian.org/cdimage/openstack/8.5.0/debian-8.5.0-openstack-amd64.qcow2

  • What size are you using and all?

Size? I've tried creating a volume from those images from 2GB to 20GB.

  • Which flavor are you using for VM creation.?

My own take on the m1.flavor:

openstack flavor create --ram 1024 --disk 10 --vcpus 1 --disk 5 m1.tiny
--
I love deadlines. I love the whooshing noise they
make as they go by.
- Douglas Adams


Mailing list:
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe :
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

--

Thanks & Regards,
Abhishek
*Cloudbyte Inc. *


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
responded Jun 21, 2016 by Abhishek_Shrivastava (3,180 points)   2 7
0 votes

On Jun 21, 2016, at 12:19 PM, Abhishek Shrivastava wrote:

​Have you tried any other flavors?

No, I never saw the point. The resources I specified was well within
the flavors rules. And the error was "Block Device Mapping is Invalid"
I can not see how changing the flavor would change that.
--
System administrators motto:
You're either invisible or in trouble.
- Unknown

responded Jun 21, 2016 by Turbo_Fredriksson (8,980 points)   7 13 19
0 votes

If it was the flavor, you would get different errors, something like
"flavor disk too small" or "out of memory". Again, I recommend to
launch an instance on local disk to see if that is working, then fix
the iscsi issue to be able to create volumes at all, first empty
volumes, then from an image and so on.

Zitat von Turbo Fredriksson turbo@bayour.com:

On Jun 21, 2016, at 12:19 PM, Abhishek Shrivastava wrote:

​Have you tried any other flavors?

No, I never saw the point. The resources I specified was well within
the flavors rules. And the error was "Block Device Mapping is Invalid"
I can not see how changing the flavor would change that.
--
System administrators motto:
You're either invisible or in trouble.
- Unknown


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

--
Eugen Block voice : +49-40-559 51 75
NDE Netzdesign und -entwicklung AG fax : +49-40-559 51 77
Postfach 61 03 15
D-22423 Hamburg e-mail : eblock@nde.ag

     Vorsitzende des Aufsichtsrates: Angelika Mozdzen
       Sitz und Registergericht: Hamburg, HRB 90934
               Vorstand: Jens-U. Mozdzen
                USt-IdNr. DE 814 013 983


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
responded Jun 21, 2016 by Eugen_Block (3,740 points)   2 2
0 votes

Hi,

First of all, think dis question did no get answered:

-I'll try that thanx. How do you do that with the "openstack" command?

If not, the command is: openstack volume create --size (size in GB) --image
(image name or id) volume_name

Just for info the cinder command was not exact, it should be: cinder create
--image-id * *--display-name

I agree with Eugen that you should make sure you can create a volume and
attach to a VM to help understand what your problem is.
This guide explains about ephemeral storage options:
https://platform9.com/support/openstack-tutorial-storage-options-and-use-cases/

By default you should be able to create VMs with ephemeral disks (not
cinder one).
Usually you can specify the directory where VM instances disks will be
stored in the compute node on nova.conf option 'instances_path' in
[DEFAULT] session. By default it should point to
'/var/lib/nova/instances/'. It is default option so, even if it is not
there, this should work.
Nova compute config options:
http://docs.openstack.org/liberty/config-reference/content/list-of-compute-config-options.html

The command to create the VM with an ephemeral disk (nova local storage and
not cinder) is:
openstack server create --image (image id or name) --flavor (flavor id or
name) vm_name

Concerning the flavor, I think the flavor you use should have the same disk
size as the disk. At least, for me when I try to boot a VM from a volume
that is not the same size of the flavor, I get BadRequest error.

Let us know if you manage to boot a VM so you can try to attach a volume to
it.

Good luck with all that.

Kind regards,
Cynthia

2016-06-21 13:36 GMT+01:00 Eugen Block eblock@nde.ag:

If it was the flavor, you would get different errors, something like
"flavor disk too small" or "out of memory". Again, I recommend to launch an
instance on local disk to see if that is working, then fix the iscsi issue
to be able to create volumes at all, first empty volumes, then from an
image and so on.

Zitat von Turbo Fredriksson turbo@bayour.com:

On Jun 21, 2016, at 12:19 PM, Abhishek Shrivastava wrote:

​Have you tried any other flavors?
>

No, I never saw the point. The resources I specified was well within
the flavors rules. And the error was "Block Device Mapping is Invalid"
I can not see how changing the flavor would change that.
--
System administrators motto:
You're either invisible or in trouble.
- Unknown


Mailing list:
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe :
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

--
Eugen Block voice : +49-40-559 51 75
NDE Netzdesign und -entwicklung AG fax : +49-40-559 51 77
Postfach 61 03 15
D-22423 Hamburg e-mail : eblock@nde.ag

    Vorsitzende des Aufsichtsrates: Angelika Mozdzen
      Sitz und Registergericht: Hamburg, HRB 90934
              Vorstand: Jens-U. Mozdzen
               USt-IdNr. DE 814 013 983


Mailing list:
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe :
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
responded Jun 21, 2016 by Cynthia_Lopes (800 points)   1
0 votes

On Jun 17, 2016, at 2:05 PM, Turbo Fredriksson wrote:

On Jun 17, 2016, at 1:12 PM, Eugen Block wrote:

Have you nova-compute.logs?

They don't say a thing, so I'm guessing it never gets
that far.

Running EVERYTHING with debugging, insensitive logging etc etc,
I noticed that Nova could not authenticate "something" (I just got
the non-descriptive "Something, something needs authentication").
I spent a whole day checking, triple checking etc. Everything WAS
ok! I'm almost sure of it! As sure I can get without fully knowing
what I'm doing at least :).

I decided that the easiest way to solve this (which I was going to
do anyway, I was just hoping to put it of until everything was working)
was to create individual service accounts for everything.

Now I can't see the Compute node any more :(.

Running "openstack --debug flavor list" (etc, etc) gives me
(with using my admin-openrc file which is supposed to give me
admin rights):

----- s n i p -----
[..]
Auth plugin password selected
authtype: password
Using auth plugin: password
Using parameters {'username': 'admin', 'project
name': 'admin', 'authurl': 'http://control:35357/v3', 'userdomainname': 'default', 'password': '***', 'projectdomainname': 'default'}
Get auth
ref
REQ: curl -g -i -X GET http://control:35357/v3 -H "Accept: application/json" -H "User-Agent: python-openstackclient keystoneauth1/2.4.1 python-requests/2.10.0 CPython/2.7.12rc1"
Starting new HTTP connection (1): control
"GET /v3 HTTP/1.1" 200 260
RESP: [200] Vary: X-Auth-Token Content-Type: application/json Content-Length: 260 X-Openstack-Request-Id: req-168f79a9-53d5-482f-841c-d9a68dbb270e Date: Tue, 21 Jun 2016 15:49:27 GMT Connection: keep-alive
RESP BODY: {"version": {"status": "stable", "updated": "2016-04-04T00:00:00Z", "media-types": [{"base": "application/json", "type": "application/vnd.openstack.identity-v3+json"}], "id": "v3.6", "links": [{"href": "http://control:35357/v3/", "rel": "self"}]}}

Making authentication request to http://control:35357/v3/auth/tokens
"POST /v3/auth/tokens HTTP/1.1" 201 11701
run(Namespace(all=False, columns=[], formatter='table', limit=None, long=False, marker=None, maxwidth=0, noindent=False, public=True, quotemode='nonnumeric'))
Instantiating compute client for VAPI Version Major: 2, Minor: 0
Making authentication request to http://control:35357/v3/auth/tokens
"POST /v3/auth/tokens HTTP/1.1" 201 11701
REQ: curl -g -i -X GET http://10.0.4.1:8774/v2/1857a7b08b8046038005b98e8b238843/flavors/detail -H "User-Agent: python-novaclient" -H "Accept: application/json" -H "X-Auth-Token: {SHA1}e3b5968af44686e0d3abfbf6e3934d6991235c46"
Starting new HTTP connection (1): 10.0.4.1
"GET /v2/1857a7b08b8046038005b98e8b238843/flavors/detail HTTP/1.1" 503 170
RESP: [503] Content-Length: 170 Content-Type: application/json; charset=UTF-8 X-Compute-Request-Id: req-c40a135f-2445-4d68-a6aa-0c37d05f363c Date: Tue, 21 Jun 2016 15:49:29 GMT Connection: keep-alive
RESP BODY: {"message": "The server is currently unavailable. Please try again at a later time.

\n\n\n", "code": "503 Service Unavailable", "title": "Service Unavailable"}
[..]
----- s n i p -----

And the web GUI gives me:

Error: Unable to get network agents info.
Error: Unable to get nova services list.
Error: Unable to get cinder services list.
Error: Unable to get Orchestration service list.

and the list of "Compute Services" is empty..

Here it's trying to connect (from what I've figured out) the compute
node. This IS up and running (on 10.0.4.3) but it seems like it haven't
(successfully) registered itself to the controller.

This is the Compute node:

----- s n i p -----
bladeA03b:~# rgrep -E '^admin|^#_tenant_|^#.*domain' /etc/nova | egrep -v '.orig|~:' | sed "s@(adminpassword = ).*@\1SECRET@" | less
/etc/nova/nova.conf:admin_username = ironic # The [ironic] section:
/etc/nova/nova.conf:admin_password = SECRET
/etc/nova/nova.conf:admin_tenant_name = service
/etc/nova/nova.conf:admin_user = nova # The [keystone_authtoken] section:
/etc/nova/nova.conf:admin_password = SECRET
/etc/nova/nova.conf:admin_tenant_name = service
/etc/nova/nova.conf:#default_domain_id =
/etc/nova/nova.conf:#default_domain_name =
/etc/nova/nova.conf:#project_domain_id =
/etc/nova/nova.conf:#user_domain_id =
/etc/nova/nova.conf:#user_domain_name =
----- s n i p -----

On the Control:

----- s n i p -----
bladeA01b:~# rgrep -E '^admin|^#_tenant_|^#.*domain' /etc/{nova,keystone,ironic} | egrep -v '.orig|~:' | sed "s@(.*(password\|token) = ).*@\1SECRET@"/etc/nova/nova.conf:admin_user = nova
/etc/nova/nova.conf:admin_password = SECRET
/etc/nova/nova.conf:admin_tenant_name = service
/etc/nova/nova.conf:#default_domain_id =
/etc/nova/nova.conf:#default_domain_name =
/etc/nova/nova.conf:#project_domain_id =
/etc/keystone/keystone.conf:admin_token = SECRET
/etc/keystone/keystone.conf:#federated_domain_name = Federated/etc/keystone/keystone.conf:#default_domain_id = default
/etc/keystone/keystone.conf:#admin_project_domain_name =
----- s n i p -----

Also, basically the only thing i can do is list users etc:

----- s n i p -----
bladeA01b:~# openstack user list
+----------------------------------+------------+
| ID | Name |
+----------------------------------+------------+
| 010049f831d84b19827ae27b72c406f1 | magnum |
| 0b7e5b0653084efdad5d67b66f2cf949 | admin |
| 0bc0163659864511a1610ba784d9e4b3 | mistral |
| 25cc2c5cf61c46329489e68656676ee4 | aodh |
| 4cf009b2dc7c4622b7230ad27f8242fe | nova |
| 4d1f0fd8c7524b7797d823eeba85cb03 | glance |
| 55f3968618b540b2a070ef845eb0c947 | ironic |
| 56e8666f2b044577934f9707ad29da5f | heat |
| 5eda7ede1be44745abd7d7815a85d927 | manila |
| 6e69a71d41da453893769ebf597bf914 | zaqar |
| 8a6694f8dde2497bbe230fbf4382f37d | trove |
| 964a9e06be3e411f9bfa80e9ea07e986 | senlin |
| a5bb89f8bbeb43d496e54109d11b1be6 | cinder |
| c0853dac1d1c4c7294f3bdfa05731c37 | barbican |
| c1bafcd2a72c429dbbf0bde8b35abb38 | murano |
| c63ad4ff853b4b72a70d64dee7aa596b | ceilometer |
| de4b432c9c7b4f1785fd600fc22df6b4 | demo |
| e298427fe3734640bfd0c6e043e13763 | neutron |
| e8bbf36bae5b4d9bb1649395b5a49886 | designate |
+----------------------------------+------------+
bladeA01b:~# openstack user list --project service

bladeA01b:~# openstack user show magnum
+--------------------+----------------------------------+
| Field | Value |
+--------------------+----------------------------------+
| defaultprojectid | f491fbef5f1748cc8fefed046973974e |
| domainid | default |
| enabled | True |
| id | 010049f831d84b19827ae27b72c406f1 |
| name | magnum |
+--------------------+----------------------------------+
bladeA01b:~# openstack project show f491fbef5f1748cc8fefed046973974e
+-------------+----------------------------------+
| Field | Value |
+-------------+----------------------------------+
| description | Default Debian service project |
| domain
id | default |
| enabled | True |
| id | f491fbef5f1748cc8fefed046973974e |
| isdomain | False |
| name | service |
| parent
id | default |
+-------------+----------------------------------+
----- s n i p -----

What "worries" me a little is that the "user list --project"
output is empty! I know that part worked once, on another
install, when I didn't use individual accounts for each
service. But the "user show" seems to indicate that the user
IS in the correct project after all..

So what is the correct way to have services authenticate themselves?
What variable/setting am I missing (or have used when I shouldn't)?

I can't see anything in the logs, even with debugging and verbose
enabled.


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
responded Jun 22, 2016 by Turbo_Fredriksson (8,980 points)   7 13 19
0 votes

On Jun 22, 2016, at 8:17 PM, Turbo Fredriksson wrote:

I can't see anything in the logs, even with debugging and verbose
enabled.

What I do see is this:

----- s n i p -----
2016-06-22 20:25:58.102 2942 DEBUG nova.service [req-e5132bf8-d3a1-4214-8cba-94ea06dfc273 - - - - -] Join ServiceGroup membership for this service compute start /usr/lib/python2.7/dist-packages/nova/service.py:236
2016-06-22 20:25:58.102 2942 DEBUG nova.servicegroup.drivers.db [req-e5132bf8-d3a1-4214-8cba-94ea06dfc273 - - - - -] DB_Driver: join new ServiceGroup member bladeA03b to the compute group, service = <nova.service.Service object at 0x7f8aa3b8e1d0> join /usr/lib/python2.7/dist-ackages/nova/servicegroup/drivers/db.py:48
----- s n i p -----

I'm not sure what it means though.


Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
responded Jun 22, 2016 by Turbo_Fredriksson (8,980 points)   7 13 19
0 votes

On Jun 22, 2016, at 8:30 PM, Turbo Fredriksson wrote:

What I do see is this:

Yay! I managed to solve a problem on my own!! :D
Mostly by guessing and hoping, but the problem was solved none
the less :D

Tweaking tenant values etc apparently solved the problem.
I now see my Compute node. And everything seems to be back where
I left of two days ago.

I now return you back to the original problem..
--
Imagine you're an idiot and then imagine you're in
the government. Oh, sorry. Now I'm repeating myself
- Mark Twain

responded Jun 22, 2016 by Turbo_Fredriksson (8,980 points)   7 13 19
0 votes

Now that my authentication problems seems to be fixed, it's back on track with
trying to boot my first instance..

On Jun 21, 2016, at 3:17 PM, Cynthia Lopes wrote:

If not, the command is: openstack volume create --size (size in GB) --image
(image name or id) volume_name

Just for info the cinder command was not exact, it should be: cinder create
--image-id * *--display-name

Thanx.

I agree with Eugen that you should make sure you can create a volume and
attach to a VM to help understand what your problem is.

Ok, so I created an empty, bootable volume. Worked just fine it seems.

I then used that when creating the instance (from Horizon).

Still the same error - Block Device Mapping is Invalid.

----- s n i p -----
bladeA01b:~# openstack volume list
+--------------------------------------+--------------+-----------+------+-------------+
| ID | Display Name | Status | Size | Attached to |
+--------------------------------------+--------------+-----------+------+-------------+
| c16975ad-dd45-41d7-b0a9-cbd0849f80e4 | test | available | 5 | |
+--------------------------------------+--------------+-----------+------+-------------+
bladeA01b:~# openstack volume show test
+--------------------------------+--------------------------------------+
| Field | Value |
+--------------------------------+--------------------------------------+
| attachments | [] |
| availabilityzone | nova |
| bootable | true |
| consistencygroup
id | None |
| createdat | 2016-06-22T20:48:31.000000 |
| description | |
| encrypted | False |
| id | c16975ad-dd45-41d7-b0a9-cbd0849f80e4 |
| migration
status | None |
| multiattach | False |
| name | test |
| os-vol-host-attr:host | bladeA01b@lvm#LVM_iSCSI |
| os-vol-mig-status-attr:migstat | None |
| os-vol-mig-status-attr:nameid | None |
| os-vol-tenant-attr:tenant
id | 2985b96e27f048cd92a18db0dd03aa23 |
| properties | |
| replicationstatus | disabled |
| size | 5 |
| snapshot
id | None |
| sourcevolid | None |
| status | available |
| type | None |
| updated
at | 2016-06-22T20:48:48.000000 |
| user_id | 0b7e5b0653084efdad5d67b66f2cf949 |
+--------------------------------+--------------------------------------+
----- s n i p -----

If I understand you correctly, this is a Cinder volume, right? Because of
the "@lvm.." part?

How can I create a local volume?

Looking under "System Information -> Block Storage Services" I see only
Cinder services..

----- s n i p -----
Name Host Zone Status State Last Updated
cinder-backup bladeA01b nova Enabled Up 0 minutes
cinder-scheduler bladeA01b nova Enabled Up 0 minutes
cinder-volume bladeA01b@lvm nova Enabled Up 0 minutes
cinder-volume bladeA01b@nfs nova Enabled Down 4 hours, 13 minutes
----- s n i p -----

This guide explains about ephemeral storage options:
https://platform9.com/support/openstack-tutorial-storage-options-and-use-cases/

Thanx, I've read something similar so I'm aware of the differences and
what they do. This one I'm going to read in more detail, because it HAD
more detail! :)

Usually you can specify the directory where VM instances disks will be
stored in the compute node on nova.conf option 'instances_path' in
[DEFAULT] session.

It was commented out, but just for the sake of it I un-commented it..

Thanx. That was actually halfway to actually be "documentation". I'll bookmark
that.

The command to create the VM with an ephemeral disk (nova local storage and
not cinder) is:
openstack server create --image (image id or name) --flavor (flavor id or
name) vm_name

----- s n i p -----
bladeA01b:/var/tmp# wget --quiet http://download.cirros-cloud.net/0.3.4/cirros-0.3.4-x86_64-disk.img
bladeA01b:/var/tmp# openstack image create --public --protected --disk-format qcow2 \

    --container-format docker --property architecture=x86_64 \
    --property hypervisor_type=docker \
    --file cirros-0.3.4-x86_64-disk.img cirros

+------------------+------------------------------------------------------+
| Field | Value |
+------------------+------------------------------------------------------+
| checksum | ee1eca47dc88f4879d8a229cc70a07c6 |
| containerformat | docker |
| created
at | 2016-06-22T21:23:03Z |
| diskformat | qcow2 |
| file | /v2/images/d4d913c3-21f3-4e7d-932c-2cb35c8131e8/file |
| id | d4d913c3-21f3-4e7d-932c-2cb35c8131e8 |
| min
disk | 0 |
| minram | 0 |
| name | cirros |
| owner | 2985b96e27f048cd92a18db0dd03aa23 |
| properties | architecture='x86
64', hypervisortype='docker' |
| protected | True |
| schema | /v2/schemas/image |
| size | 13287936 |
| status | active |
| tags | |
| updated
at | 2016-06-22T21:23:04Z |
| virtualsize | None |
| visibility | public |
+------------------+------------------------------------------------------+
bladeA01b:/var/tmp# openstack server create --image cirros --flavor m1.tiny test3
Multiple possible networks found, use a Network ID to be more specific. (HTTP 409) (Request-ID: req-381a6df8-cd8b-474a-89c4-8a5935b3d7f8)
bladeA01b:/var/tmp# openstack network list
+--------------------------------------+------------+--------------------------------------+
| ID | Name | Subnets |
+--------------------------------------+------------+--------------------------------------+
| fb1a3653-44d9-4f98-a357-c87406a8ea47 | physical | 5e3ea098-975d-460c-b313-61c11b2175d3 |
| 2bb7b8e2-188f-4e46-bf4d-ef5ec81ddb4d | network-99 | 6ef5d993-2796-4adf-a724-eae5f5d1cc53 |
+--------------------------------------+------------+--------------------------------------+
bladeA01b:/var/tmp# openstack server create --image cirros --flavor m1.tiny --nic net-id=2bb7b8e2-188f-4e46-bf4d-ef5ec81ddb4d test3
+--------------------------------------+------------------------------------------------+
| Field | Value |
+--------------------------------------+------------------------------------------------+
| OS-DCF:diskConfig | MANUAL |
| OS-EXT-AZ:availability
zone | nova |
| OS-EXT-SRV-ATTR:host | None |
| OS-EXT-SRV-ATTR:hypervisorhostname | None |
| OS-EXT-SRV-ATTR:instance
name | instance-00000003 |
| OS-EXT-STS:powerstate | 0 |
| OS-EXT-STS:task
state | scheduling |
| OS-EXT-STS:vmstate | building |
| OS-SRV-USG:launched
at | None |
| OS-SRV-USG:terminatedat | None |
| accessIPv4 | |
| accessIPv6 | |
| addresses | |
| adminPass | whateversecret |
| config
drive | |
| created | 2016-06-22T21:26:55Z |
| flavor | m1.tiny (5936ba55-7d76-4b80-8b3a-73b458b306f2) |
| hostId | |
| id | 860613fe-3834-4f72-909b-5fb4b7ff2932 |
| image | cirros (d4d913c3-21f3-4e7d-932c-2cb35c8131e8) |
| keyname | None |
| name | test3 |
| os-extended-volumes:volumes
attached | [] |
| progress | 0 |
| projectid | 2985b96e27f048cd92a18db0dd03aa23 |
| properties | |
| security
groups | [{u'name': u'default'}] |
| status | BUILD |
| updated | 2016-06-22T21:26:55Z |
| userid | 0b7e5b0653084efdad5d67b66f2cf949 |
+--------------------------------------+------------------------------------------------+
[waited a little while]
bladeA01b:/var/tmp# openstack server show test3 | grep fault
| fault | {u'message': u'Build of instance 860613fe-3834-4f72-909b-5fb4b7ff2932 aborted: Cannot load repository file: Connection to glance host http://10.0.4.3:9292 failed: Error finding address for http://10.0.4.3:9292/v1/images/d4d913c3-21f3-4e7d-932c-2cb35c8131e8: HTTPConnecti', u'code': 500, u'details': u' File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 1926, in _do
buildandruninstance\n filterproperties)\n File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 2083, in buildandruninstance\n \'create.error\', fault=e)\n File "/usr/lib/python2.7/dist-packages/osloutils/excutils.py", line 221, in exit\n self.forcereraise()\n File "/usr/lib/python2.7/dist-packages/osloutils/excutils.py", line 197, in forcereraise\n six.reraise(self.type_, self.value, self.tb)\n File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 2067, in buildandruninstance\n instance=instance)\n File "/usr/lib/python2.7/contextlib.py", line 35, in exit\n self.gen.throw(type, value, traceback)\n File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 2244, in buildresources\n reason=six.text_type(exc))\n', u'created': u'2016-06-22T21:27:28Z'} |
----- s n i p -----

Ok, that's different! I'm not running Glance on my Compute, only on my Control.

Which of these should I run on the Compute and which one on the Control?

The documentation (one of many I follow: http://docs.openstack.org/draft/install-guide-debconf/common/get_started_image_service.html) doesn't say. Only which ones to install
on the Control.

----- s n i p -----
bladeA03b:/etc/nova# apt-cache search glance | grep ^glance
glance - OpenStack Image Registry and Delivery Service - Daemons
glance-api - OpenStack Image Registry and Delivery Service - API server
glance-common - OpenStack Image Registry and Delivery Service - common files
glance-glare - OpenStack Artifacts - API server
glance-registry - OpenStack Image Registry and Delivery Service - registry server
----- s n i p -----

Currently, I have all of them only on the Control..

Concerning the flavor, I think the flavor you use should have the same disk
size as the disk.

Ok, I'll keep that in mind, thanx.

Now, this might be a stupid question, but it actually only occurred to me just now
when I looking at that missing net error. I haven't really setup my network, just
"winged" it. I' pretty sure it's not even close to working (I need to do more
studying in the matter - I still don't have a clue about how things is supposed
to work in/on the OpenStack side of things).

I've postponed it because I desperately need ANY success story - creating an
instance, even if it won't technically work would help a lot in that. I figured
it should at least TRY to start.. And I ASUME (!!) that as long as the Control
can talk to the Compute and "tell" it what to do (such as "attach this volume/image"),
it should at least be able to be created. I'm guessing the networking (Neutron)
in OS is for the instance, not for administration etc. Or, did I misunderstood
(the little I've read and actually understood about it :)?
--
Att tänka innan man talar, är som att torka sig i röven
innan man skiter.
- Arne Anka

responded Jun 22, 2016 by Turbo_Fredriksson (8,980 points)   7 13 19
...