Ask Your Question

reynoni's profile - activity

2019-04-02 05:01:04 -0500 received badge  Famous Question (source)
2019-04-02 05:01:04 -0500 received badge  Notable Question (source)
2019-04-02 05:01:04 -0500 received badge  Popular Question (source)
2019-02-19 13:51:35 -0500 commented answer no ping to Instance from neutron-gateway/0 - error "Destination Host Unreachable"

Pretty sure it's a neutron misconfiguration in openstack. I haven't had any success with conjure-up, but juju deploy openstack-base seems pretty reliable if you can manage that.

2019-02-15 12:42:42 -0500 received badge  Necromancer (source)
2019-02-14 11:15:40 -0500 commented answer no ping to Instance from neutron-gateway/0 - error "Destination Host Unreachable"

Can you verify that the router is configured with the proper upstream gateway to access the internet? If you can't ping the router, I'm guessing it isn't accessible from your network.

2019-02-14 10:56:17 -0500 asked a question Unable to retrieve volume limit information

Openstack Queens Ceph backed volumes, Maas, Juju deployed openstack-telemetry

Users are getting the following error message on logging in as well as not being able to see the limits.

Administrator could replicate. Refreshing the page typically makes the error go away

image description

/var/log/apache2/error.log from openstack-dashboard/0

[Thu Feb 14 13:56:24.098238 2019] [wsgi:error] [pid 29099:tid 140490546222848] [remote 10.50.1.195:53302] Login successful for user "admin" using domain "admin_domain", remote address 10.50.1.195.
[Thu Feb 14 13:56:25.608079 2019] [wsgi:error] [pid 29097:tid 140489893332736] Recoverable error: Service Unavailable (HTTP 503)
[Thu Feb 14 13:56:25.799720 2019] [wsgi:error] [pid 29097:tid 140490546231040] [remote 10.50.1.195:53302] Recoverable error: local variable 'limits' referenced before assignment
[Thu Feb 14 13:56:26.074376 2019] [wsgi:error] [pid 29097:tid 140490546231040] [remote 10.50.1.195:53302] No handlers could be found for logger "oslo_policy.policy"
[Thu Feb 14 13:56:38.202256 2019] [wsgi:error] [pid 29097:tid 140490546231040] [remote 10.50.1.195:53302] Call to list enabled services failed. This is likely due to a problem communicating with the Cinder endpoint. Volume Group Snapshot panel will not be displayed.
[Thu Feb 14 13:56:46.115199 2019] [wsgi:error] [pid 29096:tid 140489890182912] Recoverable error: Service Unavailable (HTTP 503)
[Thu Feb 14 13:56:46.646103 2019] [wsgi:error] [pid 29096:tid 140490621781760] [remote 10.50.1.195:53302] Recoverable error: local variable 'limits' referenced before assignment
[Thu Feb 14 13:56:46.880799 2019] [wsgi:error] [pid 29096:tid 140490621781760] [remote 10.50.1.195:53302] No handlers could be found for logger "oslo_policy.policy"
[Thu Feb 14 13:56:52.918415 2019] [wsgi:error] [pid 29098:tid 140490621781760] [remote 10.50.1.195:53516] No handlers could be found for logger "oslo_policy.policy"
[Thu Feb 14 13:56:52.942876 2019] [wsgi:error] [pid 29098:tid 140490621781760] [remote 10.50.1.195:53516] Call to list enabled services failed. This is likely due to a problem communicating with the Cinder endpoint. Consistency Group panel will not be displayed.
2019-01-18 10:20:27 -0500 received badge  Notable Question (source)
2019-01-18 10:20:27 -0500 received badge  Popular Question (source)
2018-12-20 11:25:00 -0500 commented question OpenStack doesn't apply default configurations anymore

Have you tried removing the extra physical interface mapping to verify that it is responsible for the issues? What are you trying to do that has you adding a secondary physical provider interface?

2018-12-19 07:47:18 -0500 commented answer Hypervisor summary shows incorrect total storage

If that helps or answers the question, make sure to mark it so in the post for future references.

2018-12-18 13:45:19 -0500 commented answer Hypervisor summary shows incorrect total storage

So to be clear, the hypervisor will create ephemeral disks even if it doesn't use them because that's what the flavor tells it to do. Then it registers that it created an ephemeral disk and add its storage space to it's total which is not your SAN instances.

2018-12-18 13:44:09 -0500 answered a question Self service and Provider Networks in parallel

It is possible, but it is not necessary, you can give floating public IPs from a self-service network, you can also do NAT from your router to allow traffic to pass to particular systems from reaching out to a public IP.

Stacking this technology even higher and adding layers like this will inevitably make your systems harder to manage and harder to maintain. Talking to my networking guy, he agrees that while it is probably possible, its like driving a car with your feet, sure you can do it, but why?

Use the one that does everything you need and choose a self-service network.

2018-12-18 13:32:38 -0500 commented question ceilometer problem

One of the systems it's trying to connect to isn't responding properly, what is it reaching out to, is that system running and accessible, is port 8041 open on that system?

2018-12-18 13:30:46 -0500 commented answer Hypervisor summary shows incorrect total storage

Right, by default, the instance flavors have an ephemeral disk size that is non-zero, like yours, this creates a volume for the instance on the compute node SEPERATE from your SAN. Your SAN is probably working fine and that ephemeral disk might not even be used.

2018-12-18 12:43:22 -0500 commented answer Hypervisor summary shows incorrect total storage

Okay. I think I get the disconnect. The local disk space it shows under hypervisors shows the storage that the compute hosts have on them locally, these are used by the ephemeral disks NOT your SAN backed instances. If you don't want to use ephemeral disks, you can set them to 0 like stated above.

2018-12-18 10:34:32 -0500 answered a question Hypervisor summary shows incorrect total storage

What are the Ephemeral Disk sizes on the flavors you are using? From my understanding, openstack compute uses ephemeral disks on the compute node to start the instance, then uses the storage device designated as the drive.

I've got 0 GB as my ephemeral disk size and I show 0/9.7TB on my hypervisor screen.

2018-12-12 15:14:25 -0500 commented question [Now: BUILD] openstack server show provider-instance: status ERROR

How did you install? Looks like the nova database wasn't initialized properly, claiming that the "Host 'localhost' is not mapped to any cell". Try to run this on the controller node: su -s /bin/sh -c "nova-manage cell_v2 discover_hosts --verbose" nova

2018-12-10 09:11:34 -0500 answered a question I want to deploy Openstack in Prod. but I don't want to use Rocky.

I'm currently running Rocky in production with (knock on wood) no issues. I would recommend Rocky, it fixed several issues that were plaguing me with Queens.

2018-12-06 07:57:05 -0500 commented question As you would see from the logs especially from the compute node log marked in bold, you would find that there is some kind of version mismatch going on with placement and the nova-compute service on the compute node.

Did you run the placement/nova db update/upgrade commands?

2018-12-05 09:04:21 -0500 received badge  Enthusiast
2018-12-04 10:15:44 -0500 commented question As you would see from the logs especially from the compute node log marked in bold, you would find that there is some kind of version mismatch going on with placement and the nova-compute service on the compute node.

Pretty sure openstack is saying there aren't any available nova-compute nodes. Can you run 'openstack compute service list' and let me know if there are any entries for nova-compute that show that they can be used?

2018-11-28 09:49:19 -0500 commented answer Openstack with MAAS + juju network setup

Can you answer my questions in the comment on your question?

2018-11-19 08:43:05 -0500 commented answer Failed instance launch

So when you run nova-manage cell_v2 list_hosts on the controller node, your compute nodes show up? but when you run openstack compute service list they do not show up?

2018-11-19 07:31:35 -0500 commented question nova service compute node error verifying operation

Now that I'm looking at it, the compute nodes may not be able to register because the conductor/scheduler services are not running. Can you try to resolve those services not running on the controller node?

2018-11-19 07:27:25 -0500 commented question nova service compute node error verifying operation

Try doing a service nova-compute stop waiting a few seconds then service nova-compute start. I have run into issues where it tries to auto-restart too many times, or the services restart module isn't working properly. And if you could add journalctl -xe for more information.

2018-11-19 07:25:24 -0500 commented answer Failed instance launch

Did you run su -s /bin/sh -c "nova-manage cell_v2 discover_hosts --verbose" nova on the controller node to discover the compute nodes and add them to the database?

2018-11-19 07:07:36 -0500 commented question nova service compute node error verifying operation

Does the command I posted show any specific error reasons? Alternatively, you could do service nova-compute status. Good to restart it, but there should be some information either in the status command or the logs, and your logs looked mostly empty.

2018-11-19 06:51:37 -0500 received badge  Critic (source)
2018-11-19 06:51:36 -0500 commented question Failed instance launch

Are your compute nodes up and running? Run a systemctl status nova-compute on the node and see if it's even running. Most times it isn't and you need to resolve whatever is preventing it from running.

2018-11-19 06:51:36 -0500 received badge  Commentator
2018-11-19 06:50:00 -0500 commented question nova service compute node error verifying operation

What result do you get from systemctl status nova-compute on the compute node?

2018-11-19 06:45:36 -0500 commented answer Ubuntu OpenStack wrong interface

Please submit a new question for an answer, commenting on an existing question like this with a new question is not the best or proper way to do this.

2018-11-19 06:43:53 -0500 commented answer Error verify operation with networking option 1

Can you run a systemctl status {service-name} for whichever ones are supposed to be running on that machine, but isn't? For example, systemctl status linuxbridge-agent and see if there is an error there?

2018-11-16 08:29:21 -0500 answered a question Error in the installation of nova compute service on the controller node

This is an issue with the /etc/nova/nova.conf being misconfigured. The final line states exactly where the issue is down to the line 8897 after resolving that, you should be all set on that issue.

2018-11-16 08:22:40 -0500 answered a question Ubuntu OpenStack wrong interface

I just saw your note about IP table rules. Did you make these rules persistent before the reboot? I'm pretty they get erased on a reboot. Try adding the rules back in and see if it works. I've noticed that most of the openstack auto-deployment tools do try to use eth0/eth1 rather than eno, and so I have started configuring all systems to use the eth0/eth1 naming convention. If that's not possible, there should be flags in the conjure-up process that allow you to set eno as the taget instead of eth.

Look to my comment on the question to use MAAS to set the interfaces properly.

2018-11-16 08:18:45 -0500 answered a question Error verify operation with networking option 1

I'm making a lot of assumptions since there isn't too much information. I'm assuming you are using ubuntu 18.04, and openstack rocky, following this installation guide for neutron provider 1 network on a controller node: https://docs.openstack.org/neutron/ro...

The first thing it tells you to do is apt install a bunch of packages, make sure all these packages are installed, if so, you probably need to configure one of the several config files like novainfinite posted above.

The guide posted above will walk you through each individual configuration you need to do. If you have any specific questions or can provide more information on which service isn't registering and whether or not it is properly installed, I can try to help guide you through any further steps.

2018-11-16 07:51:20 -0500 commented question Error verify operation with networking option 1

Are you doing a manual install? If so. you need to install and enable the rest of the services you want to use, like openvswitch or linuxbridge.

2018-11-15 08:08:00 -0500 commented question can not create snapshot

A snapshot from a flavor? Are you doing instance snapshots or volume snapshots? Can you track down the relevant logs?

2018-11-14 13:25:12 -0500 received badge  Teacher (source)
2018-11-14 12:50:42 -0500 answered a question Access a server with floating ip through another project

Do you have static routes set up between the subnets? I think it will be very difficult for your subnets to communicate especially if they share IP addresses.

For example, if VM1 IP was 192.168.12.50 in it's subnet, and VM2 was the same however in it's own subnet, how would the router ever know which one you are trying to route to?

So first off, I'd change one of the subnets to a slightly different value, 192.168.13.0/24 if it's available.

Through the openstack network interface, share the 192.168.12.0/24 network and create a port on the 192.168.13.0/24 subnet. After the port has been created, find the IP of the port and add a static route to the 192.168.12.0/24 that signals for the traffic intended for the .12. network goes through the port created and found above.

This is how we allow inter-project access through openstack. You will also probably have to set up security groups to allow access unless you have your instances wide open.

2018-11-14 12:41:44 -0500 answered a question migrate lvm to ceph

Do you already have a ceph cluster installed and operational? And a ceph pool for the instances? If so, with the pool accessible through cinder-volume, you can just designate where the volume is stored explicitly.

Here's some documentation with commands to run assuming you are using Rocky: https://docs.openstack.org/cinder/latest/admin/blockstorage-volume-migration.html (https://docs.openstack.org/cinder/lat...)

2018-11-14 12:36:04 -0500 received badge  Editor (source)
2018-11-14 12:35:42 -0500 edited question Self-service User Password Change - NotAuthorized

Juju deployed openstack-telemetry bundle on MaaS.

Openstack functions as normal except when any user (including domain_admin) tried to change their own password through the horizon dashboard, they get an unauthorized access error message: "Error: Unauthorized: Unable to change password." and "Error: Unauthorized. Please try logging in again."

The current web page (horizon/settings/password/) displays: "You are not authorized to access this page Login"

Horizon dashboard /var/log/apache2/error.log:

[Tue Nov 13 16:04:49.511064 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560] Unauthorized: The request you have made requires authentication. (HTTP 401) (Request-ID: req-d8991522-8179-4c8c-9aa2-c591b0df1c74)

[Tue Nov 13 16:04:49.511095 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560] Traceback (most recent call last):

[Tue Nov 13 16:04:49.511099 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]   File "/usr/lib/python2.7/dist-packages/openstack_dashboard/dashboards/settings/password/forms.py", line 67, in handle

[Tue Nov 13 16:04:49.511102 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]     data['new_password'])

[Tue Nov 13 16:04:49.511105 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]   File "/usr/lib/python2.7/dist-packages/openstack_dashboard/api/keystone.py", line 573, in user_update_own_password

[Tue Nov 13 16:04:49.511109 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]     return client.users.update_password(origpassword, password)

[Tue Nov 13 16:04:49.511112 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]   File "/usr/lib/python2.7/dist-packages/keystoneclient/v3/users.py", line 226, in update_password

[Tue Nov 13 16:04:49.511115 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]     return self._update(base_url, params, method='POST', log=False)

[Tue Nov 13 16:04:49.511118 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]   File "/usr/lib/python2.7/dist-packages/keystoneclient/base.py", line 228, in _update

[Tue Nov 13 16:04:49.511121 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]     **kwargs)

[Tue Nov 13 16:04:49.511123 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]   File "/usr/lib/python2.7/dist-packages/keystoneauth1/adapter.py", line 334, in post

[Tue Nov 13 16:04:49.511126 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]     return self.request(url, 'POST', **kwargs)

[Tue Nov 13 16:04:49.511129 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]   File "/usr/lib/python2.7/dist-packages/keystoneauth1/adapter.py", line 487, in request

[Tue Nov 13 16:04:49.511145 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1.195:54560]     resp = super(LegacyJsonAdapter, self).request(*args, **kwargs)

[Tue Nov 13 16:04:49.511148 2018] [wsgi:error] [pid 2483:tid 139843927574272] [remote 10.50.1 ...
(more)
2018-11-14 12:20:46 -0500 answered a question A communication error occurred: Connection refused

For clarity and future readers, the solution was to allow traffic across port 80 through a security group within openstack which can be done from the GUI or CLI.

2018-11-14 12:18:44 -0500 answered a question Error in dashboard service installation

Have you modified your /etc/openstack-dashboard/local_settings.py file? If so, I'd say the issue is probably in a misconfigured line in that file that these other scripts are pulling from. Verify it's integrity or load from another version.