Ask Your Question

Bernd Bausch's profile - activity

2020-02-24 21:48:38 -0600 commented question Magnum, Murano, Sahara service doesn't not supported?

I don't know whether Kolla supports them, but generally they are "supported" in the sense of "tested" and "released to production". Find documentation under docs.openstack.org/magnum, docs.openstack.org/sahara and docs.openstack.org/murano.

2020-02-24 17:01:40 -0600 commented question Architecture of keystone, swift and memcached together

You want to upload a file to a different account? I don't know if this works. Probably not, as you have to provide an authentication token that matches the account.

I suggest you create a new question.

2020-02-24 08:30:07 -0600 commented question Architecture of keystone, swift and memcached together

I don't know much but assume it will solve itself once the replicas are recreated.

2020-02-24 08:29:09 -0600 answered a question why is it that in neutron, there is a qdhcp namespace created for a dhcp enabled network, but , no namespace created for a dhcp disabled network ?

Network isolation is achieved with the ML2 type driver (assuming that ML2 is the core plugin). Either VLANs or one of the tunnel technologies GRE, VXLAN or Geneve.

DHCP namespaces don't separate tenant networks. Their purpose is to have several network interfaces with identical IP addresses on the same controller.

2020-02-24 07:19:09 -0600 commented question Architecture of keystone, swift and memcached together

1- I think so.

2- I would not bother separating three nodes. Three hundred, yes.

3- I guess it only reports that replicas are missing. I'd expect this to happen when the disk is offline.

2020-02-24 04:56:03 -0600 commented question Architecture of keystone, swift and memcached together

Yes you can have servers dedicated for objects and other servers dedicated for accounts and containers. I suppose you would run the ring builder commands with different sets of IPs.

2020-02-23 15:02:49 -0600 commented question Architecture of keystone, swift and memcached together

2- I think you have to write your own authentication solution. Perhaps https://docs.openstack.org/swift/late... helps.

4- https://stackoverflow.com/questions/1...

2020-02-23 14:35:24 -0600 commented answer proxy-server configuration file parameters
  1. My guess is that it is not defined what happens when you specify both

  2. I don't know, but most likely not in a database but inside some process' memory

2020-02-23 08:01:53 -0600 answered a question proxy-server configuration file parameters

I am not 100% sure if I understand all details correctly, but I am confident that my explanation is correct at a high level.

Keystone middleware is a Python program used by other components of OpenStack to perform authentication. It validates tokens.

You need to understand how authentication works. E.g. when you log on to the GUI, it obtains a token for you. At each action, e.g. launching an instance, listing volumes, creating a router, the same token is submitted to Nova, Cinder and Neutron. It needs to be validated by talking to Keystone.

Rather than validating the same token again and again, the keystone middleware caches it, normally in memcached. Swift has its own token cache, and the swift.cache option tells the keystone middleware to use the Swift cache instead of memcached.

If you use memcached as the caching method, Swift's token caching is done separately from the caching of other OpenStack services.

To go back to the original question, you set these options in the hope that sharing the token cache with Swift will be beneficial for performance.

2020-02-23 05:50:21 -0600 commented question proxy-server configuration file parameters
2020-02-22 05:06:28 -0600 commented question Initial Setup of DevStack Networking - Cannot ping/ssh

I don't think "private IP" is a Neutron term. "static IP" is better. It's an address from a Neutron network that was created by a user, a.k.a "tenant network". "external network" is a Neutron network that corresponds to a network outside the cloud. Floating IPs are obtained from external networks.

2020-02-22 02:54:37 -0600 commented question Initial Setup of DevStack Networking - Cannot ping/ssh

It doesn't work because there is no route to 172.16.10.4. I don't know why the VM's floating IP doesn't come out of the floating range, but I would remove the FLOATING_RANGE option and set up a new stack. Ideally on a fresh OS installation.

2020-02-21 23:29:12 -0600 commented question Initial Setup of DevStack Networking - Cannot ping/ssh

Your public-subnet is 172.24.10.0/24, and the floating IP is 172.16.10.4? How is that possible?

You ping from where?

Note that 172.16.0.0/16 is a private address range that doesn't work on the internet. By default, Devstack uses it for a fake external network.

2020-02-21 06:07:42 -0600 commented answer can we migrate created instance from openstack to opennubela??

I am grateful for the opportunity to learn. For one thing, I had thought that no other open-source clouds remained besides OpenStack. Good to see that this is not the case.

Step 4 is rather significant. Thanks for pointing it out.

2020-02-21 06:03:39 -0600 commented question Initial Setup of DevStack Networking - Cannot ping/ssh

I suppose you can't associate a floating IP because the instance is connected to the public network. Remember that floating IPs are taken from public-subnet - and your instance has such an address already!

2020-02-21 06:02:12 -0600 commented question Initial Setup of DevStack Networking - Cannot ping/ssh

It seems to me that you launched your instance on the public network, which most probably doesn't have a dhcp server. Thus, the instance doesn't get an IP address.

You can check for dhcp with openstack network show public. You can also check the instance's console log if it received an address.

2020-02-21 05:57:41 -0600 commented question Is it possible to tie nova compute AZ to cinder AZ?

Can you share the precise command for launching your instance and the precise error message?

2020-02-20 15:32:09 -0600 commented question How to correctly enable SFC plugin in RDO Packstack All-in-One deploy?

It's probably not enabled by default. Check if python-networking-sfc is installed; if not, install it and try it.

2020-02-19 16:49:59 -0600 answered a question can we migrate created instance from openstack to opennubela??

Assuming the instance uses ephemeral storage, and that OpenNebula knows what to do with qcow2, I would try this:

  1. take a snapshot from the instance
  2. download the snapshot image to a qcow2 file
  3. use that file in the OpenNebula cloud to launch an instance
2020-02-17 22:26:00 -0600 answered a question VMware vSphere on top of openstack

It just means that instances are implemented by virtual machines in a vsphere cluster. Why would you want that? Perhaps you have a vsphere installation and want to integrate it with OpenStack services such as Magnum (managing Kubernetes clusters), Sahara (Big Data), Trove (database servers) etc. Or you plan to write applications using the OpenStack API. Or you want to gradually move from Vmware to OpenStack. A few use cases by Vmware itself: https://www.vmware.com/products/opens....

2020-02-17 22:16:29 -0600 commented question Failed to populate Neutron database ( Train)

What did you do when the error occurred?

2020-02-17 07:28:51 -0600 commented question Timeout when reading response headers from daemon process 'placement-api': /usr/bin/placement-api

Perhaps it's configured to log to stderr, in which case you will probably find the logs in the systemd journal.

Config guide: https://docs.openstack.org/placement/....

2020-02-17 05:37:17 -0600 commented question Timeout when reading response headers from daemon process 'placement-api': /usr/bin/placement-api

No. This file contains messages generated by Apache. Where your Placement log is depends on your configuration; probably somewhere under /var/log.

2020-02-17 05:35:32 -0600 commented answer Error while launching instance on openstack

What do you mean by "old image in OpenStack"? There is no Cirros image in OpenStack. All images are on https://download.cirros-cloud.net/.

2020-02-17 00:24:05 -0600 commented question line 554: echo: write error: Broken pipe

There is no screenshot.

I suppose the line number refers to one of the shell scripts that are launched by stack.sh, perhaps stack.sh itself. The script name should be mentioned.

2020-02-16 17:48:11 -0600 commented question How do I Resolve Glance Cloud images Uploading Error Below to OpenStack?
2020-02-16 17:46:50 -0600 answered a question How do I solve "Missing value auth-url required for auth plugin password" error while trying to upload a cirros glance image into openstack?

The error message on the command line means that you didn't tell the openstack client how to contact Keystone. There are two ways to do that: With a command line option --os-auth-url, or an environment variable OS_AUTH_URL. See https://docs.openstack.org/python-ope....

On your controller, the environment variable can be set by sourcing admin-openrc or demo-openrc.

Furthermore, the error in the Apache log file indicates that your web server is unable to contact Glance. I guess the Glance API process is not running at that very moment.

I also think that the two messages are unrelated. The error log contains a request to list images, not to create an image. Does the timestamp in the Apache log correlate with the image create command above?

2020-02-16 06:12:34 -0600 commented answer Error while launching instance on openstack

May I delete logical volume or extend it?

If you delete cinder-volumes-pool, Cinder won't work. You can extend it, but there is no need, as nothing has been allocated from the pool so far. It has almost 20GB free space.

2020-02-16 06:09:57 -0600 commented answer Error while launching instance on openstack

I created a cinder volume without loop device on /dev/sdb1 below

No. You added an LVM (not Cinder) physical volume to the cinder-volumes volume group. What happens when you try openstack volume create --size 1 newvol? Does it work? If not, any errors in Cinder logs?

2020-02-15 19:57:51 -0600 commented question oslo_privsep.daemon.FailedToDropPrivileges

privsep-helper can't be found. It is either not installed or not in the execution PATH.

2020-02-14 21:35:58 -0600 commented question oslo_privsep.daemon.FailedToDropPrivileges

I suggest you edit your question and format the log messages so that they become readable. Use the button labeled "101010".

2020-02-14 08:26:23 -0600 commented question Timeout when reading response headers from daemon process 'placement-api': /usr/bin/placement-api

Seems the Placement service is not running. Check the placement logs for reasons why.

2020-02-13 18:17:12 -0600 commented answer how to share the image from one project to the other without making the image public

I don't know this either. Use the documentation web site and/or the help feature of the openstack client.

2020-02-13 08:14:08 -0600 answered a question how to share the image from one project to the other without making the image public
2020-02-13 05:03:51 -0600 commented answer Error while launching instance on openstack

There might be a problem with the loop device, yes, but I can't see anything wrong in the pvs/vgs/lvs output. Again, find out which command failed.

What does losetup say?

Can you create Cinder volumes manually?

Can you create logical volumes in the cinder-volumes from cinder-volumes-pool?

2020-02-13 04:46:21 -0600 commented answer Error while launching instance on openstack

I doubt that launching an instance without volume leads to an error message that the volume can't be created. In other words, the BUILDAbortException you found in the Nova log is not related to the instance without volume.

2020-02-12 17:46:56 -0600 commented answer Instance creation failing everytime

If you don't have the log anymore, it's hard to troubleshoot a past problem. I would provoke the error again, this time taking care to save the log messages.