Cloudlet on ubuntu 16.04

[tbharath91]

Hi,

I am trying to setup cloudlet locally. But in the github README it was recommended to use only ubuntu 1404 and moreover it seems if we run cloudlet inside a VM, it doesn't work. Do we have any workaround? Or do we have cloudlet packages supporting ubuntu 16.04?

[teiszler]

@tbharath91,

Currently the elijah-openstack project is bound to the OpenStack Kilo version. Kilo is not supported beyond 14.04 which is why we are also bound to it. We are working to port the OpenStack++ code to a newer version of OpenStack (Newton/Pike) which does support 16.04. There is still some testing to do to ensure that everything is working as we need to change to use neutron networking instead of the legacy nova networking, but we should hopefully be there soon.

You should however be able to run the Python command-line tools that are part of the elijah-provisioning repository in both 14.04 and 16.04. This code sits below the OpenStack integration layer and has the same functionality that is available in elijah-openstack (only without the hooks into the OpenStack framework).

Regards,
Tom

[zetelight]

Hi all, I'm sorry to ask a naive question but I'd like to double-check my understanding here.
Since @teiszler said "Currently the elijah-openstack project is bound to the OpenStack Kilo version", any latest version of OpenStack, such as Queens, will not be supported either. Am I right?
I am an absolute beginner and I cannot use DevStack(it removed the kilo branch), and Kilo release is no longer supported by the community, which means I cannot find the official installation guide, so it really takes time to install OpenStack Kilo in 2018.
Thank you for your patience and understanding!
Best,
Adam

[teiszler]

@zetelight,

You are correct. The OpenStack++ currently only works with OpenStack Kilo and devstack no longer supports Kilo. However, you needn't install it manually. I already went through the pain of distilling the Kilo installation guide into a set of Ansible tasks and so the playbook included in the elijah-openstack repository will install elijah-provisioning, OpenStack Kilo, and the OpenStack++ extensions. The README of the elijah-openstack repo has steps to perform the installation.

Hope this helps,
Tom

[sbw]

@teiszler, did you manage to port openstack++ to the later versions of Ubuntu 16.04?

[teiszler]

Hi @sbw,

Unfortunately we have not had the bandwidth to focus on porting the OpenStack extensions beyond Kilo. There is some good work that was done by folks at inWinStack to move the code to a newer OpenStack release (Newton) but it has never been fully integrated and tested.

Regards,
Tom

[sbw]

Hi @teiszler,

I have decided to use ubuntu 14.04 but when I run ansible-playbook openstackpp.yaml -i ./hosts I get stuck at TASK [openstack-controller : (NOVA) check if glance user is admin in service project] ***. How can this be corrected?

TASK [openstack-controller : (NOVA) check if nova user is admin in service project] ***
fatal: [localhost]: FAILED! => {"changed": true, "cmd": ["openstack", "--os-token", "changeme", "--os-url", "http://localhost:35357/v2.0", "role", "list", "--project", "service", "--user", "nova"], "delta": "0:00:00.451720", "end": "2019-07-23 19:34:02.047838", "msg": "non-zero return code", "rc": 1, "start": "2019-07-23 19:34:01.596118", "stderr": "ERROR: openstack No tenant with a name or ID of 'service' exists.", "stderr_lines": ["ERROR: openstack No tenant with a name or ID of 'service' exists."], "stdout": "", "stdout_lines": []}

PLAY RECAP ************************************************************************
localhost : ok=50 changed=16 unreachable=0 failed=1 skipped=24 rescued=0 ignored=0

[teiszler]

@sbw,

The service project should have been created during part of the keystone tasks. Did the playbook fail earlier and perhaps not complete all the keystone tasks? You can run a particular tagged section, for instance Keystone, by specifying --tags keystone when you launch the playbook.Try running it for just keystone to see that all of the projects/users are created properly and then you should be able to run the rest of the playbook.

Regards,
Tom

[sbw]

Hi @teiszler,
When I run ansible-playbook openstackpp.yaml -i ./hosts --tags keystone, some of the tasks were skipped especially the tasks for service project. How can I make sure that they are not skipped.

This the output:

PLAY [controller, compute] *****************************************************

PLAY [controller, compute] *****************************************************

PLAY [controller] **************************************************************

TASK [Gathering Facts] *********************************************************
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) transfer keystonedb SQL script] ********
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) check if keystone db already exists] ***
changed: [localhost]

TASK [openstack-controller : (KEYSTONE) create keystone db] ********************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) disable automatic startup of keystone] * **
changed: [localhost]

TASK [openstack-controller : (KEYSTONE) install keystone...] *******************
[DEPRECATION WARNING]: Invoking "apt" only once while using a loop via
squash_actions is deprecated. Instead of using a loop to supply multiple items
and specifying name: "{{ item }}", please use name: ['keystone'] and remove
the loop. This feature will be removed in version 2.11. Deprecation warnings
can be disabled by setting deprecation_warnings=False in ansible.cfg.
ok: [localhost] => (item=[u'keystone'])

TASK [openstack-controller : (KEYSTONE) install apache/memcached] *************
ok: [localhost] => (item=apache2)
ok: [localhost] => (item=libapache2-mod-wsgi)
ok: [localhost] => (item=memcached)
ok: [localhost] => (item=python-memcache)

TASK [openstack-controller : (KEYSTONE) replace keystone conf] *****************
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) populate identity service database] ****
changed: [localhost]

TASK [openstack-controller : (KEYSTONE) replace servername in apache conf] *****
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) create wsgi-keystone.conf] *************
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) enable identity service virtual hosts] * **
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) create wsgi dir] ***********************
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) copy keystone wsgi components] *********
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) copy keystone wsgi components] *********
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) stop keystone service if running (it wil l restart with apache2)] ***
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) remove SQLite] *************************
ok: [localhost]

TASK [openstack-controller : (KEYSTONE) wait 5s for keystone to come back up] ** *
changed: [localhost]

TASK [openstack-controller : (KEYSTONE) create identity service] ***************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) create identity endpoint] **************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) create admin project] ******************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) create service project] ****************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) create admin user] *********************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) create admin role] *********************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) create user role] **********************
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) add admin role to admin user] **********
skipping: [localhost]

TASK [openstack-controller : (KEYSTONE) create OpenStack client script] ********
ok: [localhost]
[WARNING]: flush_handlers task does not support when conditional

PLAY [compute] *****************************************************************

PLAY [controller] **************************************************************
[WARNING]: Reset is not implemented for this connection

PLAY RECAP *********************************************************************
localhost : ok=18 changed=4 unreachable=0 failed=0 s kipped=9 rescued=0 ignored=0

[sbw]

@teiszler,

I run the whole script again and it was successful.

However, I can not succeed with creating the private network.

$ source ~/admin-openrc.sh
-su: /opt/stack/admin-openrc.sh: No such file or directory

[teiszler]

When the playbook runs, it runs as root (and subsequently sets the permissions for this script to be owned by root) so it should be under /root/admin-openrc.sh. It appears that whatever user you are currently, the home directory is /opt/stack. Can you sudo -i to become root?

[sbw]

Thank you @teiszler for your gudance.

However, I can not log into the dashboard, even after editing the password in source ~/admin-openrc.sh

export OS_PROJECT_DOMAIN_ID=default
export OS_USER_DOMAIN_ID=default
export OS_PROJECT_NAME=admin
export OS_TENANT_NAME=admin
export OS_USERNAME=admin
export OS_PASSWORD=changeme
export OS_AUTH_URL=http://localhost:35357/v3

[teiszler]

Changing admin-openrc.sh will have no effect on the admin user's password as it will have already been set in the database. You would only need to change admin-openrc.sh if you changed your password in the dashboard (or via the openstack client). If you didn't change keystone_admin_pw in elijah-openstack/ansible/roles/openstack-controller/vars/main.yml, then it should be the default. If you did change it before you ran the playbook, then you should be able to log in with that password and admin as the user. Failing that, you could try to recreate the keystone database by manually running the SQL script that the playbook runs:

cat ~/keystonedb.sql | mysql --user root --password={{mysql_pw}}

Of course you will still have to know the mysql password to do this. Also, since this will drop the table and recreate it, and keystone is a dependency for the rest of the services, I would caution against this. It is likely that you would have to re-run the entire playbook (definitely you would have to re-run the keystone tasks) so that all the endpoints that get stored in keystone for the other services gets repopulated.

Another way to test, is by using the openstack client. If you source admin-openrc.sh and then do some nova commands, it uses the same authentication credentials. Assuming you were able to create the private network with this method, then those credentials are the ones that should also be used to log into the dashboard as the administrator.

Hope that helps,
Regards,
Tom

[sbw]

Hello @teiszler,

Following your first comment, I have not changed admin-openrc.sh and I also haven't changed keystone_admin_pw in elijah-openstack/ansible/roles/openstack-controller/vars/main.yml. I still can't login.

[teiszler]

If you didn't change it, then it should be admin/changeme according to what is in GitHub. But you said "even after editing the password in source ~/admin-openrc.sh" which lead me to believe that you changed the password in admin-openrc.sh. Can you still do nova commands at the commandline? If you can, then authentication may not be the problem preventing you to log in. It could be a problem with horizon (the dashboard component) trying to connect to keystone. You may have to start looking at the logs in /var/log/keystone to figure out what the true error is.

[sbw]

Thank you @teiszler for your prompt response. I have successfully logged in. However, there is an option of updating the password. I am going to change it.

[sbw]

@teiszler,

I want to install the Cloudlet Gateway usisng the Ansible playbook whilst specifying the gateway tag with command:
ansible-playbook -i ./hosts openstackpp.yaml --tags gateway

I get this error. How can I go about it?
ansible-playbook -i ./hosts openstackpp.yaml --tags gateway
ERROR! the playbook: openstackpp.yaml could not be found

[teiszler]

@sbw,

It looks like your cwd (current working directory) isn't where openstackpp.yaml is. You need to cd to elijah-provisioning/ansible/ where the playbook lives and run it from there. Or provide the full path instead of a relative one when you tell it to launch the playbook.

Regards,
Tom

[sbw]

@teiszler, I cd to elijah-openstack /ansible and the error was resolved. Thank you.

However, I can not proceed with ansible-playbook -i ./hosts openstackpp.yaml --tags gateway. Am stuck at TASK [cloudlet-gateway : (GATEWAY) install dependecies]. It seems there is a problem with my network.

failed: [localhost] (item=[u'openvpn', u'easy-rsa', u'python-dev', u'libssl-dev', u'nodejs', u'npm']) => {"ansible_loop_var": "item", "cache_update_time": 1564143292, "cache_updated": false, "changed": false, "item": ["openvpn", "easy-rsa", "python-dev", "libssl-dev", "nodejs", "npm"], "msg": "'/usr/bin/apt-get -y -o \"Dpkg::Options::=--force-confdef\" -o \"Dpkg::Options::=--force-confold\" install 'openvpn' 'easy-rsa' 'libssl-dev' 'nodejs' 'npm'' failed: E: There are problems and -y was used without --force-yes\n", "rc": 100, "stderr": "E: There are problems and -y was used without --force-yes\n", "stderr_lines": ["E: There are problems and -y was used without --force-yes"], "stdout": "Reading package lists...\nBuilding dependency tree...\nReading state information...\nThe following extra packages will be installed:\n gyp libc-ares-dev libc-ares2 libccid libjs-node-uuid liblzo2-2\n libpkcs11-helper1 libssl-doc libv8-3.14-dev libv8-3.14.5 node-abbrev\n node-ansi node-archy node-async node-block-stream node-combined-stream\n node-cookie-jar node-delayed-stream node-forever-agent node-form-data\n node-fstream node-fstream-ignore node-github-url-from-git node-glob\n node-graceful-fs node-gyp node-inherits node-ini node-json-stringify-safe\n node-lockfile node-lru-cache node-mime node-minimatch node-mkdirp\n node-mute-stream node-node-uuid node-nopt node-normalize-package-data\n node-npmlog node-once node-osenv node-qs node-read node-read-package-json\n node-request node-retry node-rimraf node-semver node-sha node-sigmund\n node-slide node-tar node-tunnel-agent node-which nodejs-dev opensc pcscd\n zlib1g-dev\nSuggested packages:\n pcmciautils node-hawk node-aws-sign node-oauth-sign node-http-signature\n systemd\nThe following NEW packages will be installed:\n easy-rsa gyp libc-ares-dev libc-ares2 libccid libjs-node-uuid liblzo2-2\n libpkcs11-helper1 libssl-dev libssl-doc libv8-3.14-dev libv8-3.14.5\n node-abbrev node-ansi node-archy node-async node-block-stream\n node-combined-stream node-cookie-jar node-delayed-stream node-forever-agent\n node-form-data node-fstream node-fstream-ignore node-github-url-from-git\n node-glob node-graceful-fs node-gyp node-inherits node-ini\n node-json-stringify-safe node-lockfile node-lru-cache node-mime\n node-minimatch node-mkdirp node-mute-stream node-node-uuid node-nopt\n node-normalize-package-data node-npmlog node-once node-osenv node-qs\n node-read node-read-package-json node-request node-retry node-rimraf\n node-semver node-sha node-sigmund node-slide node-tar node-tunnel-agent\n node-which nodejs nodejs-dev npm opensc openvpn pcscd zlib1g-dev\n0 upgraded, 63 newly installed, 0 to remove and 3 not upgraded.\nNeed to get 442 kB/6860 kB of archives.\nAfter this operation, 27.0 MB of additional disk space will be used.\nWARNING: The following packages cannot be authenticated!\n liblzo2-2 libc-ares2 zlib1g-dev libssl-dev libssl-doc nodejs pcscd opensc\n openvpn libc-ares-dev nodejs-dev\n", "stdout_lines": ["Reading package lists...", "Building dependency tree...", "Reading state information...", "The following extra packages will be installed:", " gyp libc-ares-dev libc-ares2 libccid libjs-node-uuid liblzo2-2", " libpkcs11-helper1 libssl-doc libv8-3.14-dev libv8-3.14.5 node-abbrev", " node-ansi node-archy node-async node-block-stream node-combined-stream", " node-cookie-jar node-delayed-stream node-forever-agent node-form-data", " node-fstream node-fstream-ignore node-github-url-from-git node-glob", " node-graceful-fs node-gyp node-inherits node-ini node-json-stringify-safe", " node-lockfile node-lru-cache node-mime node-minimatch node-mkdirp", " node-mute-stream node-node-uuid node-nopt node-normalize-package-data", " node-npmlog node-once node-osenv node-qs node-read node-read-package-json", " node-request node-retry node-rimraf node-semver node-sha node-sigmund", " node-slide node-tar node-tunnel-agent node-which nodejs-dev opensc pcscd", " zlib1g-dev", "Suggested packages:", " pcmciautils node-hawk node-aws-sign node-oauth-sign node-http-signature", " systemd", "The following NEW packages will be installed:", " easy-rsa gyp libc-ares-dev libc-ares2 libccid libjs-node-uuid liblzo2-2", " libpkcs11-helper1 libssl-dev libssl-doc libv8-3.14-dev libv8-3.14.5", " node-abbrev node-ansi node-archy node-async node-block-stream", " node-combined-stream node-cookie-jar node-delayed-stream node-forever-agent", " node-form-data node-fstream node-fstream-ignore node-github-url-from-git", " node-glob node-graceful-fs node-gyp node-inherits node-ini", " node-json-stringify-safe node-lockfile node-lru-cache node-mime", " node-minimatch node-mkdirp node-mute-stream node-node-uuid node-nopt", " node-normalize-package-data node-npmlog node-once node-osenv node-qs", " node-read node-read-package-json node-request node-retry node-rimraf", " node-semver node-sha node-sigmund node-slide node-tar node-tunnel-agent", " node-which nodejs nodejs-dev npm opensc openvpn pcscd zlib1g-dev", "0 upgraded, 63 newly installed, 0 to remove and 3 not upgraded.", "Need to get 442 kB/6860 kB of archives.", "After this operation, 27.0 MB of additional disk space will be used.", "WARNING: The following packages cannot be authenticated!", " liblzo2-2 libc-ares2 zlib1g-dev libssl-dev libssl-doc nodejs pcscd opensc", " openvpn libc-ares-dev nodejs-dev"]}

PLAY RECAP *****************************************************************************************
localhost : ok=2 changed=0 unreachable=0 failed=1 skipped=0 rescued=0 ignored=0

[teiszler]

@sbw,

I believe there is a timeout for an Ansible task and if you had a slow connection when downloading all these packages, it could have thought that the task failed even though it just wasn't finished. I would try it again, and since most of those packages should have been installed the first time around, it should finish more quickly.

The error message also indicates that you may have other package errors you need to resolve. You can try running apt-get ugprade manually and see what the error entails.

Regards,
Tom